Content uploaded by Marc Kochzius
Author content
All content in this area was uploaded by Marc Kochzius on Dec 30, 2016
Content may be subject to copyright.
Content uploaded by Marc Kochzius
Author content
All content in this area was uploaded by Marc Kochzius on Mar 31, 2016
Content may be subject to copyright.
Content uploaded by Marc Kochzius
Author content
All content in this area was uploaded by Marc Kochzius on Apr 01, 2015
Content may be subject to copyright.
Chitipothu S, Cariani A, Bertasi F, Stagioni M, Kochzius M, Blohm D, Tinti T, Landi M
(2014) Invertebrate DNA chip: opportunities and challenges in the development and
application of microarrays for marine biodiversity studies. In: Rogers JV (ed)
Microarrays: principles, applications and technologies. Genetics - research and issues.
Nova Science Publishers, Inc., pp 101-134; ISBN: 978-1-62948-713-7
Chapter 7
INVERTEBRATE DNA CHIP: OPPORTUNITIES AND
CHALLENGES IN THE DEVELOPMENT AND
APPLICATION OF MICROARRAYS FOR MARINE
BIODIVERSITY STUDIES
Srujana Chitipothu1,3, Alessia Cariani2,4∗ , Fabio Bertasi2,5,
Marco Stagioni2,4, Marc Kochzius1,6, Dietmar Blohm1,
Fausto Tinti2,4 and Monica Landi2,7
1Centre for Applied Gene Sensor Technology (CAG), FB2-UFT,
University of Bremen, Bremen, Germany.
2Department of Experimental Evolutionary Biology,
University of Bologna, Bologna, Italy.
3Current Affiliation: Vision Research Foundation, Sankara Nethralaya,
Chennai, Tamilnadu, INDIA.
4Department of Biological, Geological and Environmental Sciences, University of
Bologna, Italy.
5Institute for Environmental Protection and Research,
ISPRA Via di Castel Romano 100, 00128Rome - Italy.
6Marine Biology, Free University of Brussels (VUB), Brussels,Belgium.
7CBMA (Centre of Molecular and Environmental Biology), Department of Biology,
University of Minho, Campus de Gualtar, 4710-057 Braga, Portugal.
ABSTRACT
DNA microarrays for marine biodiversity studies have been developed and tested in
several groups, spanning different applications. They are employed to characterize
ecological communities circumventing uncertainties and challenges associated with the
conventional techniques of taxonomy employed to characterize these communities.
Examples come from identifications of fishes and phytoplankton to monitoring of
∗Corresponding author:alessia.cariani@unibo.it.
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
2
harmful algae. Although DNA microchips support a great diversity of applications and
provide a wealth of findings in functional genomics and environmental studies, limits in
ecological applications are known. So far, DNA microarrays used in biodiversity studies
provide only qualitative data in terms of presence or absence of the species and its usage
is greatly limited in providing the quantitative measurement of environmental samples.
Moreover, DNA microarrays are commonly restricted to identify only those species that
are targeted by the probes implemented on the chip. In turn, developing a broad spectrum
array has technological challenges in terms of probe designing, experimental
optimization, and statistical analysis. In this book chapter, we provide a critical
assessment on promises and pitfalls of DNA microarrays as tool for marine invertebrate
species identification. We implemented a DNA-chip prototype to identify 15 species of
marine invertebrates from European Seas, including crustaceans, molluscs, and
polychaetes, based on the two mitochondrial markers, cytochrome oxidase subunit I and
16S rRNA. Challenges involved in oligonucleotide probe design, in silico evaluation, and
difficulties encountered through hybridization experiments are here explored. Specificity
and sensitivity of the probes have also been critically evaluated to verify the suitability of
the selected markers for microarray probe design. Since ultimate application of DNA
microarray to resolve complex environmental samples is a major challenge, we made an
attempt to analyze gut contents of predator fishes. The problems encountered during this
analysis, as the presence of target and not target species that could affect the specificity
and the sensitivity of the DNA-chip to distinguish low and high abundant target species
from a background of non-targets, were explored. We also reviewed advantages and
disadvantages of DNA microarray technology compared to other molecular identification
methods that recently spread, i.e. DNA barcoding and next generation sequencing. The
outcome of the Invertebrate DNA Chip prototype served as a proof-of-concept for the
identification of selected marine invertebrates and prey species of demersal fishes by
DNA microarray. The potential of such microarrays can encompass several fields of
scientific applications in marine biodiversity and ecosystem sciences, as marine
ecosystem diversity and environmental monitoring, seafood quality control, and
understanding food webs and ecosystem functioning.
INTRODUCTION
Challenges in Marine Biodiversity and Ecosystem Sciences
Compared to terrestrial ecosystems, very little is known about marine biodiversity, and its
changes in species richness and ecosystem functioning, mostly due to sampling and
taxonomic constraints [1]. Although the situation of marine biodiversity is alarming, expertise
to assess biodiversity is currently being lost [2]. The necessary monitoring of marine
ecosystem diversity is thus hampered, due to the lack of taxonomic expertise and tools for the
identification of sampled organisms [3].
Monitoring marine biodiversity through the standard quali-quantitative analysis of fish
gut content dates back in the early 80s [4][5][6]. Since then, such analysis has been used as a
tool across several studies [7]. The stomach content analysis has been adopted as supportive
tool in several studies, ranging from fisheries assessment [8], marine food web and successful
monitoring of marine ecology [9], and monitoring of marine invertebrates and bioindicator
organisms, which are important prey species [10]. The approach has been successful to detect
abundant and taxonomically well-known species in the gut contents. However, quantitative
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
3
and qualitative constrains can be commonly identified [11][12]. Because prey is often
partially or fully digested, the diagnostic characters for species identifications are frequently
lost, and the identification of organisms is then mainly based on their remaining parts or
traces [11]. For such reason, unambiguous prey identification through taxonomic approaches
undergoes a limiting bottleneck. In addition, the standard taxonomic identification of
organisms by stereo-microscopic analysis is extraordinarily time consuming and marginally
useful for routine analysis, yielding only partial results.
Analyses of Fish Gut Contents- Traditional versus Molecular Approaches
Much of the knowledge on fish diet habits was initially based on analysis of stomach
contents involving methods of occurrence, numerical, weight, and volumetric frequencies
[6][13][14]. The advent of molecular tools employed in ecological and fisheries studies
expanded and integrated the standard taxonomic approach [15][16][17][18]. Based on the
nucleotide differences among species-specific DNA sequences, the concept of genetic
markers for taxa identification, has gained acceptance and widespread application
[19][20][21][22][23]. In fact, molecular approaches to species identification are considered as
the most efficient solution to inventory all life forms for biodiversity conservation and
monitoring [15][24]. Successful reports were published about DNA-based identification of
marine animals, such as eggs, larvae, and adults of fishes [20][25][26], zooplankton [27], and
invertebrate larvae [28]. The sequences of partial mitochondrial Cytochrome c Oxidase
subunit 1 (COI) gene along with those of the 16S rRNA (16S) and Cytochrome b (Cytb)
genes were proposed as standard tools (e.g. barcodes) for a global bio-identification system of
eukaryotes [15].
The need for simultaneous handling of large numbers of samples in mass scale studies
has led to improved molecular techniques able to perform multiple sequences analyses, such
as DNA microarray technologies [29]. Microarrays, or so-called microchips, are one of the
most powerful innovations in molecular biology since their emergence nearly 30 years ago
[30]. Allowing parallel hybridizations of hundreds to thousands of nucleic acids probes on a
small surface area, DNA microchips transformed and accelerated the technical framework of
several research disciplines. The first idea of identifying a DNA with radioactive-labelled
short single-stranded DNA molecules started in 1975 [31], where the target DNA to be
analyzed was immobilized on a membrane. After fourteen years, in 1989, Saiki et al., [30]
reversed the design and started the microarray revolution, where many DNAs could be
identified simultaneously by immobilizing many specific oligonucleotides on a solid support
(i.e. the microarray slide). Since its first implementation [32], a rapid increase in novel
technological solutions is being explored to date. In the last few decades, DNA microarrays
have evolved as a promising tool for detecting thousands of environmental samples in
biodiversity studies [33][34][35]. Some research programs have highlighted the potential of
the DNA-microarrays as high throughput molecular tools to detect marine organisms for
effective monitoring of marine ecology [18].
DNA microarrays were used to identify different taxa in the marine realm: bacteria [36],
pathogenic vibrios infecting shellfish and fishes [37][38], benthic fishes [29][39][40],
phytoplankton [41][42] and harmful algae [43][44][45][46].Invertebrates form a major part of
marine ecosystems and largely contribute to the marine food webs [47]. Being a highly
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
4
diverse ecological group of organisms, the identification of marine invertebrates at the species
level is a real challenge. Until now, DNA-DNA hybridization techniques were developed to
identify larvae of marine invertebrates [48]. Hence, one of the main objectives of this chapter
was to develop a high-throughput DNA microarray to detect marine invertebrate species in
the gut content of benthic marine fish. At the methodological level, the potential for high-
throughput and automation of microarray platforms is the critical factor driving interest
towards the development of DNA microarray technology, even though sometimes the arrays
are lacking consistency of results and involve complex process of data analysis [49].
DNA MICROARRAYS – CHALLENGES
Several types of challenges can be identified in the development and application of DNA
microarrays for environmental studies [50], compared to the highly standardized functional
genomics arrays, and to the most recently used approaches, such as the DNA barcoding
methodology [15][16] [17].
1. Biological challenges: Due to the large number of species across different phyla of
invertebrates, it is highly impossible to develop an array having probes specific for
all the organisms. Such microarray can be limited only to a selected number of
species with referenced DNA sequences available for probe designing. However, the
use of probes to detect organisms at multiple hierarchical taxonomic levels can
enhance the accuracy of the array. Developing a broad-spectrum microarray with
probes suitable for the identification of organisms at species and higher taxonomic
levels will be a task facing technological challenges in probe design and optimization
of hybridization conditions [51][52].
2. Selection of appropriate gene markers for probe design: This is a critical step, as the
candidate gene should have a nucleotide substitution rate able to discriminate even
closely related species, but also allow distinguishing individuals of the same species.
For such reasons, the mitochondrial DNA genes 16S, COI, and Cytb are widely used
markers in probe design for species identification. The COI has been widely found to
be efficient in differentiating closely related species [15][24][40][53]. On the
contrary, the 16S was more suitable for identifying inter-specific variation rather than
intra-specific variation [54]. Some organisms, such as fishes, were successfully
differentiated also using the Cytb [40].
3. Technological challenges: The selection of appropriate gene markers would require
in-silico and in-vitro standardization experiments. Major technological challenges
include 1) the design of species-specific oligonucleotide probes, 2) the amplification
of the target genes from single- or multi-target DNA template, 3) the optimization of
hybridization conditions for highest binding efficiency of target to probes, 4) the
detection of the optimum signal intensity from the fluorescent labels, 5) the
sensitivity of the array to detect target species in highly diverse environmental
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
5
samples, and finally 6) the development of standard statistical methods for data
analysis [55][56].
Invertebrate DNA Chip – Development
This chapter deals with the development and implementation of an Invertebrate Chip
(hereafter INV-CHIP) for the identification of selected species of marine invertebrates and the
critical assessment of its reliability in species identifications. The INV-CHIP targets
important prey species of demersal fishes in the Mediterranean Sea and bioindicator
organisms, such as polychaetes, that are difficult to be identified by morphological characters.
The INV-CHIP aimed at enhancing ecosystem research on energy flow, food web structure,
and environmental monitoring, for its potential to facilitate the identification of digested prey
in gut contents analysis. By increasing the accuracy of multi-target identifications and the
performance for large mass-scale biodiversity studies, the INV-CHIP could further strengthen
the regular monitoring of bioindicator organisms in marine ecological research. The
specificity and sensitivity of the INV-CHIP were tested using single and multiple targets as
well as environmental samples such as gut contents of demersal fish, whose prey species
composition was in parallel analysed by standard stereomicroscopic analysis.
MATERIALS AND METHODS
Sample Collection and DNA Extraction
A total of 267 specimens from 15 species of crustaceans (7 species), molluscs (4) and
polychaetes (4) were collected from different areas of the Adriatic Sea and North Sea (Table
1) during scientific expeditions. Species were selected on the basis of their ecological traits
and geographical range, focusing on prey of demersal fish and bioindicator organisms.
Simultaneously, samples of Chelidonichthys lucerna (tub gurnard) were collected from the
North Adriatic Sea in order to obtain suitable gut samples for testing the INV-CHIP. All the
individual specimens collected were morphologically identified at species level and stored in
ethanol 96% at 4°C. A collection of voucher specimens and tissue samples was created and
genomic DNA was extracted from each sample as previously described [29].
After recording of biological data, the entire gut and stomach content of the tub gurnard
individuals were excised and stored in ethanol 96% at -20°C. Genomic DNA extraction of
individual gut content was performed by collecting biological materials (i.e. particles and
prey remains) from the stomach. The taxonomic analysis of the species content with standard
morphological approach was performed under a stereo-microscope. The selected biological
materials were then homogenised with a pestle or a homogenizer to break down
agglomerates. An aliquot of approximately 200 µL was pelleted by centrifugation and ethanol
was removed. The pellet was dried at 37°C. The total genomic DNA extraction was
performed following the protocol from Deagle et al., [57].
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
6
Table 1. List of 15 invertebrate species and respective area of collection
Taxon
Family
Species Name
Species
Code
Collection Site
Individuals
collected
Crustacea
Alpheidae
Alpheus glaber
Ag
Adriatic Sea
10
Crangonidae
Crangon crangon
Cc
North Sea
64
Goneplacidae
Goneplax
rhomboides
Gr
Adriatic Sea
10
Portunidae
Liocarcinus
depurator
Ld
Adriatic Sea
10
Portunidae
Liocarcinus
vernalis
Lv
Adriatic Sea
10
Euphausiidae
Meganyctiphanes
norvegica
Mn
Adriatic Sea
10
Grapsidae
Pachygrapsus
marmoratus
Pm
Adriatic Sea
10
Mollusca
Octopodidae
Eledone cirrhosa
Ec
Adriatic Sea
10
Corbulidae
Lentidium
mediterraneum
Lm
Adriatic Sea
10
Veneridae
Chamelea gallina
Cg
Adriatic Sea
10
Ommastrephidae
Illex coindetii
Ic
Adriatic Sea
10
Polychaeta
Nereididae
Hediste diversicolor
Hd
Mediterranean
Sea
22
Nephtyidae
Nephtys hombergii
Nh
Mediterranean
Sea & North Sea
56
Pectinariidae
Pectinaria koreni
Pk
North Sea
13
Sigalionidae
Sigalion mathildae
Sm
North Sea
12
Amplification and Sequencing of Genetic Markers
The COI and 16S markers were targeted for amplification and sequencing. The COI and
16S were PCR amplified using the universal primer pairs LCO 1490-HCO 2198 [58], and
16SarL-16SbrH [59], respectively. All PCR products were quali-quantitatively checked by
1.5%-agarose gel electrophoresis. Exo-Sap purified PCR amplicons were sequenced using the
Big Dye terminator cycle sequencing kit (Applied Biosystem) according to the
manufacturer’s protocol. Both DNA strands were sequenced. Electropherograms were
visually inspected and edited for indels and substitutions, by using the software MEGA
version 5 [60]. After editing, the identity of the target sequences was checked by nucleotide
BLAST search against the NCBI GenBank [61].
Sequence Analysis and Capture Probe Designing
Sequences of at least two individuals of each species and additional sequences from
NCBI were used for capture probe design. COI and 16S sequences of non-target crustaceans,
molluscs, and polychaetes from European Seas were obtained from NCBI (Table 2). The
oligonucleotide probes were designed by the Chip Designing Internet Service (CDIS)
provided by the Miconet service of the Techno Mathematics Department of the University of
Bremen (Microarray Collaboration Network web page www.miconet.uni-bremen.de) [62].
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
7
The final criteria used for the 16S probe design were: the length of the oligo in the range
of 23 to 50 bp, GC content: 35-55%, specific binding with the target sequences, and
hybridization temperature in the range of 70-80°C, considering a total 1M salt concentration.
The COI capture probes were designed as for the 16S but with much narrower criteria (oligo
length from 23-35 bp and GC content of 40-50 %). The probes were checked for any
secondary structure, self-complementarities, and absence of cross-reactions against targets
and non-targets. Minimal free energy (mfe) structures were computed by using the software
RNA fold [63]. A capture probe included in the experimental tests only if it recognized the
proper target species with 0 mismatches and showed at least 3 mismatches distributed in the
center of the oligo with respect to other target and non-target species. At least two capture
probes for each marker and species were selected (Table 3).
Preparation of DNA Microarrays and Hybridization Experiments
Preparation of DNA microarray slides and hybridization experiments were carried out as
described in details in Kochzius et al. [29][40], using a set of 96 spots arranged in 4 rows and
24 columns on a glass slide. The array layout of the prototype INV-CHIP is displayed in
Figure 1.
Table 2. For each target and non target species, 16S and COI sequences used for probe
design. A: sequences newly obtained for the INV-CHIP and B: sequences retrieved from
the NCBI database
Taxon
Species
16S
COI
Targets
Non-
targets
B
Targets
Non-
targets
B
A
A
A
B
Crustacea
Alpheus glaber
4
0
601
3
0
1085
Crangon crangon
5
0
2
0
Goneplax rhomboides
3
0
3
0
Liocarcinus depurator
3
0
3
1
Liocarcinus vernalis
3
0
2
0
Meganyctiphanes
norvegica
4
1
3
2
Pachygrapsus marmoratus
4
1
2
0
Mollusca
Eledone cirrhosa
3
3
726
3
1
1624
Lentidium mediterraneum
2
0
2
0
Chamelea gallina
2
3
3
2
Illex coindetii
2
1
3
1
Polychaeta
Hediste diversicolor
4
0
162
5
0
461
Nephtys hombergii
13
0
9
0
Pectinaria koreni
3
3
4
0
Sigalion mathildae
3
0
4
0
Total
70
1489
58
3170
Table 3. List of species-specific capture probes (COI and 16S) developed and implemented in the INV-CHIP prototype.
Species Name
Probe
Name
Sequence in 5'- to 3'- direction
Length (bp)
%GC
Tm(°C)
A. glaber
Ag-16S_1
Ag-16S_2
Ag-COI_1
Ag-COI_2
AGTTAGGTCATTATGCTGGGGCG
GTTTAAGTTAGGTCATTATGCTGGGGCGG
TCCTACTCAGGCTCCCAGTTCTA TTAGGGATCTTCTCGCTACACCTCG
23
29
23
25
52
48
52
52
82,56
83,7
82,13
82,49
C. crangon
Cc-16S_1
Cc-16S_2
Cc-16S_3
Cc-COI_1
Cc-COI_2
Cc-COI_3
GAATGATCGGACAAGGGGCTAACTG
GGAATGAATGATCGGACAAGGGGCTAAC
GGAATGAATGATCGGACAAGGGGCTAACTG
CCTGCTTTAACTCTTCTTCTATCTAGAGGA
CCTCCTGCTTTAACTCTTCTTCTATCTAGAGGA
CTAGAGGATTAGTAGAAAGAGGAGTAGGAACTGGA
25
28
30
30
33
35
52
50
50
40
42
43
82,4
83,79
84,85
79,49
81,95 82,36
G. rhomboides
Gr-16S_1
Gr-COI_2
Gr-COI_1
GAACTTGTATGAATGGTTGGACAAAGGAAAAGCTGTCTCTATTATA
CGCCTCTGTTGATATAGGTATTTTTTCCCTTCATT
ACTTCCTCCCTCTTTAACTCTCTTACTAAT
46
35
30
35
37
37
83,83 82,1
79,22
L. depurator
Ld-16S_1
Ld-16S_2
Ld-COI_2
Ld-COI_1
CTTAGAAAAATGTATTGGGTTGGGGCGAC
CTTAGAAAAATGTATTGGGTTGGGGCGACTAAGGTATA
TTACTCCCACCTTCGCTAACTCTTCTCCTC
TCTACCCTCCCCTATCGGCTGCTATT
29
38
30
26
45
39
50
54
82,51 83,84
84,73 85,64
L. vernalis
Lv-16S_1
Lv-16S_2
Lv-COI_1
Lv-COI_2
ATTCAGAAAAGTGTATTGGGTTGGGGCG
AGGTTTATTCAGAAAAGTGTATTGGGTTGGGG
GCTTCTTCTCATGAGAGGTATAGTCGAGAG
TGACCCAGTTCTCTATCAGCACTTGTTTTG
28
32
30
30
46
41
47
43
84,23
82,98 81,36
82,96
M. norvegica
Mn-16S_1
Mn-16S_2
Mn-COI_2
Mn-COI_1
ATGGTCGGACAAGAAACAGACTGTCTTCA
GGTGTAGCAGCTTTAATAGAAGGTCTGTTCGAC
CCACCTTCTTTAACTCTTTTATTAGGCAGAGGTC
GGCAGAGGTCTTGTAGAAAGAGGAGTC
29
33
34
27
45
45
41
52
84,11 83,63
82,07 82,75
P. marmoratus
Pm-16S_1
Pm-16S_2
Pm-COI_1
Pm-COI_2
TCGAGGGCTATAAAGGCTTGGTG
AGTCTAATTCGAGGGCTATAAAGGCTTGGTG
ACCTCCCTCCTTATCTCTCTTACTTACAAGA
CAACATACGCTCTTATGGTATGACAATAGACCA
23
31
31
33
52
45
42
39
82,23 83,75
81,83 81,66
Species Name
Probe
Name
Sequence in 5'- to 3'- direction
Length (bp)
%GC
Tm(°C)
E. cirrhosa
Ec-16S_2
Ec-16S_1
Ec-COI_1
Ec-COI_2
GATTGGGGTGATCAAGGAATAAAAAAGA
GATAAATAAACCGAGAGTTTTGC
CCTCCCCTATCAAGTAATTTAGCCCACATAG
ACCCTCTGTTGACCTAGCAATTTTCTCTTTACATC
28
23
31
35
36
35
45
40
78,62 73,4
82,11 83,27
L. mediterraneum
Lm-16S_1
Lm-16S_3
Lm-16S_2
Lm-COI_1
Lm-COI_2
TGGGGAAAGGTATGAAGGGACTG CTGGAGGGCTAATCGAATGGAAAGT
CAGCGTAATTTTCTACTGGAGGGCT
CCCCATTATCCGGTAATACAGCTCACTC
AGACTTTCTTATTTTATCGCTACACCTCGGT
23
25
25
28
31
52
48
48
50
39
81,98 81,9
81,97 82,81
81,72
C. gallina
Cg-16S_1
Cg-COI_2
Cg-COI_1
GGAGAATGGTATGAATGGTTTAACGTAGAATAACTGTCTTTGGAA
CTTATCTAGGGCTCTGTCTCATTCGGG CTTAGGTTCTGATTGTTGCCGGTG
45
27
24
36
52
50
83,93 82,51
81,25
I. coindetii
Ic-16S_1 Ic-
COI_2 Ic-
COI_1
GCTTGAATTTTTTAAAGGGACGAGAAGACCCTAATGAGCTTATAA
CCACCATCTTTAACTATATTACTAGCCTCTTCAGC
CCCTTATCTAGAAATTTATCTCATGCTGGACCC
45
35
33
36
40
42
84,23 81,35
81,97
H. diversicolor
Hd-16S_1
Hd-16S_2
Hd-COI_1
Hd-COI_3
Hd-COI_2
GCTAATTAATCACACACACACCCAA
TGGGACAACCTAAAGACAAATAAACCTCTTAGCT
GTCCGTCAGTAGACCTTGCAATCTTC
ATGAACAGTATACCCGCCATTAGCCAGAAATATTG
GGTCTCATCTATTTTGGGAGCCCTAAAC
25
34
26
35
28
40
38
50
40
46
78,96
83,06 81,91
83,44 81,65
N. hombergii
Nh-16S_1
Nh-16S_2
Nh-COI_2
Nh-COI_1
AAAAGGAACAAGTTTGGTTGGGGC
AAAAGGAACAAGTTTGGTTGGGGCGACAAAG
CCCTTCTTTAATTCTTCTTGTTATATCCGCAGCTG
TATCCGCAGCTGTAGAAAAAGGAGTC
24
31
35
26
46
45
40
46
82,3 85,92
82,4 81,45
P. koreni
Pk-16S_2
Pk-16S_1
Pk-16S_3
Pk-COI_1
Pk-COI_2
Pk_COI_3
GGGGCGGCTGAGGAAAATTAAATC GAGGGCCAAGCTGTCTCTTTAGT
GGATCAAAGAAAATAGCTACCTCGGGG
CTTTTCTCTCCACTTAGCCGGGATC CCCATTAATACTTGCTGCTCCAGAC
CCCCTTATCAAGAAACCTTGCACATGCG
24
23
27
25
25
28
50
52
48
52
48
50
81,83 82,19
81,73 81,93
80,67 84,63
S. mathildae
Sm-16S_1
Sm-16S_2
Sm-COI_3
Sm-COI_2
Sm-COI_1
CGACCCAGGAGCATTTAAACCCT
GAAGGCTGGAATGAACGGATAAACGAG
CCTTCAGTTGATCTCGCTATCTTTTCTCTTCATATTGC
ACTCCCTCCTTCTCTAATTCTTCTTTTATCTTCA
TCTTCAAGAGCCGTTGAAAAAGGAGTT
23
27
38
34
27
52
48
39
35
41
82,79 82,23
82,94 80,73
81,92
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
10
Figure 1. Layout of the prototype INV-CHIP. Probes name as listed in Table 3; PC: positive control;
NC: negative control; Empty: empty spots.
Target Amplification - Multiplex 16S and COI PCR
In order to reduce the number of PCR reactions needed to carry out all single target
hybridizations, a multiplex PCR was optimized combining the single 16S and COI protocols.
Multiplex PCR conditions were similar to those used in single PCR, except for the annealing
temperature set to 45°C. PCR reactions were performed in 100 µL with 0.3 µM of each
primer. Target DNAs for microarray hybridization experiments were amplified using Cy5-5’-
labelled forward primers as described in Kochzius et al. [29][40].
Experiments for Testing Probe Specificity
According to the optimized hybridization conditions, a series of single target
hybridizations experiments were performed both 16S and COI markers to test the specificity
of the probes selected for each microarray. In order to check the behavior of each target
species on the chip and also the signal patterns obtained by hybridizing different markers, the
first series of hybridization experiments were performed to amplify and hybridize the two
markers separately. Single target hybridizations were realized with three individuals for each
target species and each hybridization experiment was replicated three times. According to
these results, oligos were selected for the final INV-CHIP, which included both 16S and COI
probes. A total of 64 capture probes with satisfactory behaviour were chosen for the
combined INV-CHIP. Both 16S and COI markers from all the 15 target species were
amplified in a multiplex PCR with two sets of primers as described above. The final
concentration of amplicons in the hybridization solution was 20 nM instead of 10 nM (as in
the case of single marker hybridizations) assuming that the two markers were amplified with
a comparable yield in the multiplex PCR. As all the hybridization parameters had already
been checked and the signal pattern of every single target was known, two non-interfering
targets were hybridized together on to the microarray to reduce the number and time of
experiments.
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
11
Experiments for Testing Probe Sensitivity: Binary and Ternary Mixtures
The sensitivity of the microarray was tested using a hybridization series with two targets
on the same chip, to check the reciprocal influence in term of the signal intensities given by
the probes of one target when the other one was present at different concentrations. In this
test, three sets of targets were selected from each of the three taxonomic groups represented
on the INV-CHIP: Crustacea, Mollusca, and Polychaeta. The two targets were selected
according to the criterion that, although closely related, they did not show any cross
hybridization in the single target experiments.
The sensitivity was further assessed with a similar three target hybridization series to
check the number of targets that the microarray can identify without any interference. As in
the two target experiments, three species belonging to the same three taxonomic groups were
simultaneously hybridized in the microarray. The 16S and COI markers were amplified in a
multiplex PCR. In the first experiments, all the targets were added in equal amounts, while
further experiments were carried out keeping one target at the minimum concentration and
adding two other targets in the highest concentrations. The PCR products were hybridized on
to the INV-CHIP to check if there was a correlation between the template concentration, the
PCR amplification of different targets and the signal intensities.
Experiments on Environmental Samples
The suitability of the INV-CHIP in ecological studies was assessed by analyzing the gut
content of seven individuals of the demersal fish (C. lucerna). Previous work had analyzed
the diet of several demersal fishery resources, and the preliminary results showed that C.
lucerna feed on some of the crustacean species targeted by the INV-CHIP [64]. The gut
content samples were collected, total DNA extracted as described above, and hybridized onto
the chip. Verification of the hybridization results was carried out by cloning and sequencing
the 16S and the COI amplicons obtained from the gut content samples.
Measurement of Fluorescence Signals and Data Analysis
Hybridization signals were measured as described in Kochzius et al., [40]. The
hybridization spots were replicated according to the configuration of the array (Figure 1) (n =
1 for probes, n = 14 positive controls - PC and n = 4 for negative controls - NC). The mean
and standard deviation of the pixels in the spots were generated by the software Genepix pro
v4.1 (Axon, Union City, USA), along with the local background. Local background was the
mean of the pixels measured in a ring around the spot twice its diameter.Data were pre-
processed by calculating the ratio standard deviation (SD)/mean for each spot (coefficient of
variation, CV) and excluding the signals with CV > 1 from the analysis.
Hybridization signals for all tests were analysed using the same procedure. Negative
control signals were compared with local background values by means of Welch's t test for
unpaired samples with pooled SD [65]. When significantly different at p < 0.05, the highest
among negative control- and background- signal was retained as the control (CTRL) signal. If
not significantly different, the two measurements were averaged as the CTRL signal for
subsequent analyses.
Differences between hybridization signals were tested by one-way ANOVA. Variances
were preliminarily tested for heteroscedasticity using Bartlett test [66]. When necessary, data
transformations were done to stabilize the variances. If variances could not be stabilized using
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
12
appropriate transformations, ANOVA analysis was run on the untransformed data [67]. On
significant ANOVA results, differences among groups were tested by multiple pairwise t
tests. The Benjamini–Hochberg procedure [68] was used to control the false discovery rate
and correct α values. For ANOVA analyses and pairwise tests the significance threshold was
set at p < 0.01. All statistical analyses were performed using R Project for Statistical
Computing v.2.15 [69].
RESULTS
Sequencing of the Targets
For each target species, an initial set of 10 individuals was processed for DNA extraction,
gene amplification, and sequencing. At least two 16S and COI sequences for each target
species were obtained (Table 2).
In Silico Characterization of Capture Oligonucleotides
For the 16S probe design, a series of different calculations were necessary to obtain
oligos with satisfactory features for all target species, increasing the allowed range of
thermodynamic properties. The length and GC-content parameters of the selected 16S probes
are 23–46 bp and 35–52%, and those of the COI probes are 23–35 bp and 35–54%. All the
probes selected are shown in Table 3 along with the parameters considered for selection.
Optimization of the INV-CHIP Hybridization Conditions
Single target hybridizations were carried out with three different targets at four different
hybridization temperatures (50°C, 55°C, 60°C, and 65°C), over three different hybridization
time-spans (2h, 3h, and 4h). The optimal hybridization temperature and time-span on this
microarray were 60°C and 3h respectively (data not shown). According to these results, a
series of optimization experiments was performed with the remaining 12 target species,
testing the same individual for each species at three different hybridization temperatures
(50°C, 55°C, 60°C) for 3h. The overall signal pattern was evaluated in terms of the
relationship between the number of specific signals and any false positive and false negative
signals produced by the probes. Even though the specific signal was lost from some probes
the unspecific signals were significantly reduced at 60°C, and all further experiments were
always carried out at 60°C for 3h.
Specificity of INV-CHIP
A total of 64 capture probes with satisfactory behavior were chosen for the INV-CHIP.
As all the hybridization parameters had already been checked and the signal pattern of every
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
13
single target was known, two non-interfering targets were hybridized together on to the
microarray. The overall signal pattern along with significance of the tested values against the
CTRL values are shown in Table 4. At least one probe of each marker per target species gave
a significant signal, except the 16S probes of Crangon crangon, Goneplax rhomboides and
Lentidium mediterraneum which were lost, probably due to an uneven amplification of the
two markers in the multiplex PCR reaction. Most of the specific signals were significant
considering the stringent threshold of p < 0.01, while two COI probes, Lv-COI_1 and Nh-
COI_1, gave a specific significant signal only if p < 0.05 (Table 4).
Few probes showed false positive signals (binding with non specific targets). This
occurred for the COI probes Lv-COI_2 and Nh-COI_1, and for the 16S probes Ld-16S_1 and
Ld-16S_2 which cross-hybridized between the two congeneric target species Liocarcinus
depurator and L. vernalis. Several probes did not give the expected hybridization signal when
the specific target was hybridized on the INV-CHIP, leading to false negative signals. This
negative performance of the INV-CHIP was much more relevant for the 16S probes (with 11
out of 30 probes not producing a detectable signal when compared to the CTRL), than for the
COI probes (where only five out of 34 probes provided false negative signals).
Sensibility of the INV-CHIP
Multiple target hybridizations were performed to 1) check the influence of closely related
targets upon each other when hybridized together and 2) evaluate the effect on the signal
intensities of the probes of one target when other targets are present at different
concentrations, and 3) assess the detection limit of a target in complex mixtures.Binary
mixtures were set up by selecting targets from the three main taxonomic groups considered in
this study: Chamalea gallina, Eledone cirrhosa (Mollusca),G. rhomboides, Pachyrgapsus
marmoratus (Crustacea), Nephtys hombergii, andSigalion mathildae (Polychaeta) (Figure 2).
Figure 2a shows the signal pattern obtained with the molluscs E. cirrhosa and C. gallina. In
the experiments performed, the presence of a second target did not influence significantly the
signal of specific hybridization, except for the C. gallina capture probe Cg-COI_1, whose
hybridization signal was totally lost when the E. cirrhosa target was added independently
from the amount. A gradual correspondence could be observed between the hybridization
signal intensity decrease and increase from probes of one target, when the concentration of
the second was increased or decreased; however, a clear correlation between target amount
and hybridization signal intensity was not observed.
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
14
Table 4. Summary of the hybridization results of the 15 targets on the INV-CHIP.
Probes name as listed in Table 3. **: specific signals with p < 0.01; *: specific signals
with p < 0.05; FN: False negative signals; FP: False positive signals, with associated
significance
Probes
Ag+Cc
Ec
Gr+Pk
Ld+Ic
Lm+Nh
Lv+Cg
Mn+Hd
Pm+Sm
Ag-16S_1
FN
Ag-16S_2
**
Ag-COI_1
**
Ag-COI_2
**
Cc-16S_1
FN
Cc-16S_2
FN
Cc-16S_3
FN
Cc-COI_1
**
Cc-COI_2
**
Cc-COI_3
**
Cg-16S_1
**
Cg-COI_1
FN
Cg-COI_2
**
Ec-16S_1
**
Ec-16S_2
**
Ec-COI_1
**
Ec-COI_2
**
Gr-16S_1
FN
Gr-COI_1
FN
Gr-COI_2
**
Hd-16S_1
**
Hd-16S_2
**
Hd-COI_1
**
Hd-COI_2
**
Hd-COI_3
**
Ic-16S_1
**
Ic-COI_1
**
Ic-COI_2
**
Ld-16S_1
**
FP *
Ld-16S_2
**
FP **
Ld-COI_1
**
Ld-COI_2
**
Lm-16S_1
FN
Lm-16S_2
FN
Lm-16S_3
FN
Lm-COI_1
**
Lm-COI_2
**
Lv-16S_1
**
Lv-16S_2
**
Lv-COI_1
*
Lv-COI_2
FP **
**
Mn-16S_1
FN
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
15
Probes
Ag+Cc
Ec
Gr+Pk
Ld+Ic
Lm+Nh
Lv+Cg
Mn+Hd
Pm+Sm
Mn-16S_2
**
Mn-COI_1
FN
Mn-COI_2
**
Nh-16S_1
**
Nh-16S_2
**
Nh-COI_1
FP **
*
Nh-COI_2
**
Pk-16S_1
FN
Pk-16S_2
**
Pk-16S_3
**
Pk-COI_1
**
Pk-COI_2
FN
Pk-COI_3
**
Pm-16S_1
FN
Pm-16S_2
**
Pm-COI_1
FN
Pm-COI_2
**
Sm-16S_1
**
Sm-16S_2
**
Sm-COI_1
**
Sm-COI_2
**
Sm-COI_3
**
Table 5. Checklist of preys of Chelidonichthys lucerna and relative abundance indexes,
modified from Stagioni et al, 2011. In bold species targeted by the INV-CHIP. %N,
percentage in number; %W, percentage in weight; %F, frequency of occurrence; %IRI,
percentage of index of relative importance of prey items; PSA, % prey taxon over all
preys in all predators
PREY TAXA
%N
%W
%F
%IRI
PSA
CRUSTACEA
89.69
58.01
7.47
92.24
72.51
Philocheras spp.
44.38
1.53
14.52
26.51
23.98
Philocheras bispinosus
15.1
0.96
4.75
3.04
33.5
Goneplax rhomboides
14.16
29.52
28.95
50.3
60.41
Liocarcinus spp.
7.09
4.47
10.95
5.04
59.1
Processa spp.
6.15
1.8
5.26
1.66
17.52
Liocarcinus depurator
5.96
20.32
10.02
10.47
73.7
Alpheus glaber
0.83
1.01
3.23
0.24
19.63
Solenocera membranacea
0.62
1.51
1.95
0.17
47.18
Squilla mantis
0.43
2.95
1.78
0.24
54.32
Macropodia spp.
0.28
0.16
0.34
0.01
26.02
Jaxea nocturna
0.21
0.51
0.93
0.03
16.05
Liocarcinus maculatus
0.23
0.43
0.68
0.02
33.92
Corystes cassivelaunus
0.11
1.79
0.51
0.04
81.96
Upogebia spp.
0.08
0.28
0.34
<0.01
16.83
Lophogaster typicus
0.08
0.02
0.34
<0.01
6.42
Liocarcinus vernalis
0.08
0.09
0.17
<0.01
41.35
Sicyonia carinata
0.06
0.09
0.17
<0.01
30.56
Pontophilus spinosus
0.06
0.09
0.17
<0.01
26.84
Parapenaeus longirostris
0.04
0.15
0.17
<0.01
17.35
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
16
Table 5. (Continued)
PREY TAXA
%N
%W
%F
%IRI
PSA
Brachynotus spp.
0.04
0.05
0.17
<0.01
11.9
Chlorotocus crassicornis
0.04
0.05
0.08
<0.01
51.53
Liocarcinus pusillus
0.02
0.02
0.08
<0.01
16.68
Pisidia longimana
0.02
0.02
0.08
<0.01
10.67
Ebalia granulosa
0.02
<0.01
0.08
<0.01
52.46
Processa modica
0.02
<0.01
0.08
<0.01
1.29
Munida spp.
0.02
<0.01
0.08
<0.01
100
Melicertus kerathurus
0.02
<0.01
0.08
<0.01
2.99
TELEOSTEI
6.72
39.74
1.92
7.45
74.65
Engraulis encrasicolus
0.6
11.91
2.29
1.14
81.89
Gobius niger
0.6
8.77
2.04
0.76
86.93
Lesueurigobius friesii
0.49
1.57
1.53
0.13
52.88
Cepola macrophthalma
0.13
2.01
0.59
0.05
98.84
Pachygrapsus marmoratus
0.41
0.32
1.1
0.03
46.96
Lesueurigobius suerii
0.26
0.86
0.68
0.03
38.63
Trisopterus minutus
0.08
1.8
0.17
0.01
99.78
Arnoglossus laterna
0.08
1
0.25
0.01
73.34
Gobius spp.
0.08
0.52
0.34
0.01
27.4
Serranus hepatus
0.04
0.79
0.17
0.01
87.19
Callionymus maculatus
0.08
0.35
0.25
<0.01
56.29
Gaidropsarus biscayensis
0.08
0.18
0.34
<0.01
40.43
Merlangius merlangus
0.02
0.93
0.08
<0.01
100
Deltentosteus quadrimaculatus
0.04
0.15
0.17
<0.01
92.81
Callionymus spp.
0.04
0.13
0.17
<0.01
84.73
Callionymus risso
0.04
0.11
0.17
<0.01
100
Merluccius merluccius
0.02
0.24
0.08
<0.01
100
Microchirus variegatus
0.02
0.03
0.08
<0.01
79.17
Pomatoschistus minutus
0.02
0.01
0.08
<0.01
100
MOLLUSCA
1.57
0.55
0.41
0.04
30.13
Corbula gibba
0.34
0.19
1.44
0.03
4.69
Turritella communis
0.21
0.22
0.93
0.02
9.49
Anadara demiri
0.04
0.01
0.17
<0.01
2.94
Nassarius spp.
0.06
0.01
0.25
<0.01
0.48
Arca tetragona
0.04
0.02
0.17
<0.01
3.69
Tellina spp.
0.04
0.01
0.17
<0.01
3.01
Epitonium spp.
0.02
0.03
0.08
<0.01
10.65
ANELLIDA
0.03
0.01
0.01
<0.01
4.92
Aphrodita aculeata
0.02
0.01
0.08
<0.01
3.8
Sternaspis scutata
0.02
0.01
0.08
<0.01
9.67
NOT DETERMINED
1.99
1.7
0.92
0.28
17.38
The second set of the two target hybridizations was carried out with the crustaceansG.
rhomboides and P. marmoratus (Figure 2b). According to the results obtained in the
specificity tests, the two captures Gr-16S_1 and Gr-COI_1 did not produce any detectable
signal when G. rhomboides was hybridized at standard concentration, while increasing
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
17
amounts of target yielded a significant signal with probe Gr-COI-1. Overall, all the specific
probes gave hybridization signals according to the relative increase in the target
concentration.
The third experiment was conducted with various mixtures of polychaetes N. hombergii
and S. mathildae (Figure 2c). A correlation between the signal intensity and the target
concentration was observed to a certain extent. The rapid increase in the amount of the second
target did not affect the signal of the first target and vice versa, except for the capture probe
Nh-COI_1. In the latter case, the signal was quite low compared to the other capture probes
and there was no significant signal obtained from the capture probe when the S. mathildae
target was added in higher concentrations.
Table 6. Results of the INV-CHIP analysis of seven gut contents of Chelidonichthys
lucerna, compared to those obtained by morphological identification. SP: morphological
identification at species level. GE: morphological identification at genus level. Bold text:
targets detected by morphological analysis and by the INV-CHIP. Shaded cells: target
detected by the INV-CHIP but not by morphological analysis. St118-120: sampling
station; TL1-4: individual code for each specimen collected in each sampling station
SAMPLES
Gut 1
Gut 2
Gut 3
Gut 4
Gut 5
Gut 6
Gut 7
St120
TL1
St118
TL2
St120
TL3
St118
TL3
St119
TL1
St118
TL4
St119
TL2
INV-CHIP Targets
Lentidium
mediterraneum
Chamelea gallina
Illex coindetii
Eledone cirrhosa
Meganychtiphanes
norvegica
Crangon crangon
Alpheus glaber
X(GE)
Goneplax rhomboides
X(SP)
X(SP)
Pachigrapsus
marmoratus
Liocarchinus depurator
X(GE)
X(SP)
X(GE)
X(GE)
X(GE)
Liocarcinus vernalis
X(GE)
X(GE)
X(GE)
X(GE)
Nephtys hombergii
Hediste diversicolor
Pectinaria koreni
Sigalion mathildae
Non Target species
Philocheras sp.
X
X
X
X
X
X
X
Gobidae (otolith
remains)
X
X
X
Processa sp.
X
X
Callyonimus sp.
X
X
X
Bony fish (trace)
X
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
18
Figure 2. Signal pattern of binary mixture hybridizations; a) C. gallina, E. cirrhosa (Mollusca); b) G.
rhomboides, P. marmoratus (Crustacea); c) N. hombergii, S. mathildae (Polychaeta). * p< 0.01.
To further assess the detection limit of the INV-CHIP when analyzing complex mixtures,
three target species of the same phyla were hybridized together. The molluscs C. gallina, E.
cirrhosa and Illex coindetii, the crustaceans G. rhomboides, P. marmoratus, and L. vernalis,
and the polychaetes N. hombergii, S. mathildae, and Pectinaria koreni were used in ternary
mixtures hybridization experiments (Figure 3). In order to carry out these multiple-target
experiments, three targets at different concentrations were amplified in a multiple target PCR
with two sets of primers (16S and COI). PCR products were hybridized onto the INV-CHIP
to determine whether there was any correlation between the template concentration, the PCR
amplification of different targets, and the signal detected.
The results from the experiments with the ternary mixtures of molluscan targets are
displayed in Figure 3a. Hybridization pattern showed that only the target E. cirrhosa was
properly amplified in the multiplex reactions, while the two other targets were amplified with
a minor yield. The observed signal intensity from these specific probes was very low
compared to the single target experiments, thereby exhibiting decreased significance in many
experiments. All the E. cirrhosa captures gave the expected signals when the target was
equivalent to the other two templates as well as when the concentration of the target was
lower respect to the other two. The overall signal from the C. gallina captures was very low,
with no significant signal observed from both the C. gallina COI probes. This differed from
the results of specificity tests, where the capture Cg-COI_2 gave the expected hybridization
signal. The C. gallina 16S probe pattern reflected the concentration of the targets, providing a
false negative result when the specific target was lower than the two other targets. Capture
probes targeting I. coindetii gave significant hybridization signals in most of the experiments,
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
19
with intensity values well correlated to the template concentrations in the multiplex PCRs. As
observed for C. gallina, a false negative signal was obtained from two out of three I. coindetii
probes when the specific target was at a lower concentration than the other two targets.
The second set of multi-targets consisted of three crustacean targets from G. rhomboides,
P. marmoratus,and L. vernalis. These were hybridized after amplification in a multi-target
PCR with different template proportions (Figure 3b). All the three targets were amplified as
expected; signals were obtained from at least one specific probes of each. As expected, the
two capture probes Gr-16S_1 and Gr-COI_1 did not produce any detectable signal either
when G. rhomboides was hybridized at standard concentration or increasing its amounts of
the ternary mixture. Moreover, the significant signal from probe Gr-COI_2 was not detected
when the concentration of the specific target was lowered with respect to the other two
targets.
Hybridization patterns of L. vernalis were similar to those obtained from the other two
crustacean species. Three capture probes provided the expected specific signals, while the
probe Lv-COI_1 did not produce any significant signal. The P. marmoratus capture probes
showed a higher level of performance than that observed in the specificity test series, since all
probes gave significant signals when the target was added at equal or higher concentrations.
Only the capture probes Pm-16S_1 and Pm-COI_1 produced non-significant signals when P.
marmoratus was analyzed at the lowest concentrations, confirming the expectation of false
negative signals.
The third set of ternary mixture hybridization experiments included the polychaetes N.
hombergii, S. mathildae,and P. koreni. The signal pattern indicated that all three targets and
both markers were successfully amplified in the multiplex PCR (Figure 3c). All capture
probes produced at least one significant signal when the corresponding specific target was
present in the mixture, with the exception of two P. koreni (Pk-16S_1 and Pk-COI_2) and one
N. hombergii (Nh-COI_1) capture probe that yielded false negative signals consistently with
the results of specificity tests. False negative signals were indeed given by several capture
probes when the target concentration was lowered with respect to the other two targets (N.
hombergii: all capture probes; S. mathildae: Sm-16S_2, Sm-COI_1, and Sm-COI_2; P.
koreni: Pk-16S_2). The higher number of COI probes giving false negative signals suggested
a higher susceptibility of COI probes to the variation of target amount in the mixture. Any
hybridization series gave significant correlation between target concentration and signal
intensity.
Species Identification in Environmental Samples with INV-CHIPs: Prey Identification
in Gut Contents of Chelidonichthys lucerna
After testing the INV-CHIP specificity and sensibility with artificial mixtures of known
composition, the applicability of the microarray to the taxonomic analysis of fish gut contents
was tested by the gut content analysis of tub gurnard C. lucerna (Linnaeus, 1758) individuals
collected in the Adriatic Sea [64]. In this geographic location, C. lucerna feeds on several
crustacean species that were targeted by the INV-CHIP as shown in Table 5. This evidence
represents on optimal test to assess the INV-CHIP specificity and sensitivity on field samples.
The stomach contents of each individual was analyzed under the stereomicroscope and
taxonomically recognized using identification keys, giving attention to assess the
presence/absence of the species targeted by the INV-CHIP.
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
20
Figure 3. Signal pattern of ternary mixture hybridizations a) C. gallina, E. cirrhosa and I. coindetii
(Mollusca); b) G. rhomboides, P. marmoratus and L. vernalis (Crustacea) and c) N. hombergii, S.
mathildae and P. koreni (Polychaeta). * p< 0.01.
The INV-CHIP hybridization experiments on gut contents were carried out according to
the protocol optimized with artificial mixtures. The assay was replicated at least four times
for each gut sample, due to high variability of the hybridization signals and the occurrence of
various artifacts in the INV-CHIP hybridization results, probably due to the biological
complexity of the environmental mixture.
Results were evaluated in terms of intensity and significance of hybridization signals.
The inferred presence/absence of the INV-CHIP target species was compared to the results of
the standard taxonomical analysis. The summarized INV-CHIP results were shown in Table 6
which displayed also the species identified by morphological keys, either at species level (SP)
or at genus level (GE).
Results for the gut content samples were also presented as box plot graphs in Figure 4, to
point out the great variability of the signal patterns obtained. This suggested that additional
replicates need to be performed for each samples in order to obtain more reliable results in
terms of species identification.
All targets detected by the standard taxonomical analysis were identified in parallel by
the INV-CHIP (Bold text in Table 6), but not all target detected by the INV-CHIP were
detected by the expert taxonomist under the stereo microscope (Shaded cells in Table 6). The
INV-CHIP results for the Gut3 sample were further confirmed by cloning and sequencing of
both 16S and COI amplicons. Sequences were successfully blasted to homologous sequences
of taxa belonging to Crustacea, Decapoda. Once checked against the target species sequences
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
21
obtained in this study, amplicon sequences gave the best alignment with sequences of the
target species G. rhomboides (data not shown).
CONCLUSION
The specificity of the INV-CHIP target probes was satisfactory and the INV-CHIP
appears to be an effective tool to identify target invertebrate species. Statistical analysis of the
hybridization signal intensity gave effective measures of true positive identification, while
ruling out both false positive and negative signals. The suitability of COI and 16S markers for
species discrimination with a microarray platform was tested on vertebrates [29] and
prompted us to test and compare their individual performance in discriminating the target
invertebrate species. While 16S probes specifically identified 12 out of 15 targets, the COI
capture probes gave specific signals in all 15 target species. The p-values associated to the
signal intensity showed that the COI targets bound very specifically to the capture probes on
the INV-CHIP and less false negative signals were observed with respect to 16S captures
(Table 4). This suggests the mitochondrial COI gene is a useful marker for discriminating
invertebrates at the species level on a microarray platform. The suitability of the COI gene as
species identification marker has been proven in a great variety of invertebrate taxa, such as
Copepoda [27][70], Lepidoptera [71][72], Culicidae [73], Araneae [74]. However, the
combined COI-16S INV-CHIP microarray exhibited the advantage of a more robust pattern
of hybridization signals for all 15 target species, providing internal validation of species
identification. The target specific hybridization pattern can be used as reference for assessing
the presence of the target in the analysis of unknown environmental samples (Figure 4). The
PCR multiplexing of COI and 16S targets represents the additional value of the INV-CHIP by
reducing cost and time of analysis.
The INV-CHIP specificity was also tested against annelid worm species (class
Polychaeta). The failure of hybridization of four non-target species of polychaetes
corroborated the microarray specificity as demonstrated by a weak and non-significant cross-
hybridization with a single N. hombergii probe (data not shown). However, such preliminary
validation of the INV-CHIP specificity needs to be confirmed by further experiments using a
higher number of non-target species closely related to all the 15 target species. A
“PhyloChip” model, i.e a chip carrying a hierarchical set of probes recognizing species at
multiple taxonomic levels [52], could represent a suitable approach for enhancing microarray
platform specificity. In the INV-CHIP, this approach was not accomplished because the
design of capture probes suitable for the identification of the invertebrate taxa at higher
taxonomic level was not feasible due to the unsatisfactory output generated by the CDIS
analyses.
The reliable application of the INV-CHIP to biodiversity and ecosystem monitoring
would require the simultaneous detection of multiple target species in a single experiment and
efficient discrimination of the target species even in samples with high biological complexity.
The results of the hybridization tests carried out with artificial binary and ternary mixtures of
target species in different proportions demonstrated that the hybridization of one target was
still possible in the presence of unbalanced amount of other targets. Even when a target
species was underrepresented, the INV-CHIP was able to detect significant hybridization
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
22
signals (Figure 2; Figure 3). Similar results were also obtained from the analysis of
environmental samples (i.e. fish gut contents) whose quali-quantitative composition was
actually unknown. However, the INV-CHIP cannot effectively be used for absolute, or
relative target quantification as there was no deterministic correlation between the amount of
the target in the mixtures and the intensity of the hybridization signals; however, increases in
the signal intensity was observed at higher target concentrations.
Further sensitivity tests of the INV-CHIP to detect target species within artificial
mixtures with large amount of non-target species (i.e. mimicking more complex
environmental samples) were conducted by hybridizing the PCR-amplified targets from the
mixing of genomic DNA from phytoplankton, polychaetes, and gut content of fishes. Even as
target signals gradually decreased with increasing the non-target concentrations, the target
signal could still be detected in the presence of a 20-fold background concentration (data not
shown).
Given the performance of the INV-CHIP in the specificity and sensitivity tests, results
obtained from the analysis of the unknown environmental samples appeared satisfactory. The
presence of target species in the gut contents was resolved by the INV-CHIP in the majority
of samples analysed, and the comparative morphological and molecular analyses consistently
identified most of the target invertebrate species. The gut contents analyses revealed a high
frequency of crustacean Decapoda (Table 5), that are common prey in the diet of tub gurnard
in the Adriatic Sea [64]. The positive hybridization signals given by the INV-CHIP,
indicating the presence of the target species in the gut sample, were validated either by the
standard morphological microscopic identification, or by the cloning and sequencing the
target PCR products of the gut content template DNA. COI and 16S sequences obtained from
the preliminary molecular validation test carried out on the same gut sample showed high
levels of BLASTn identity (>87%) with sequences of several crustacean Decapoda and a
100% identity with sequences of Goneplax rhomboides (data not shown).The consistency of
results given by such independent approaches further validated the high power of INV-CHIP
in identifying target invertebrate species.
Some discrepancies between morphological and molecular identification concerned
species that were detected by the hybridization experiments, but not by the standard
taxonomical stereo-microscopical analysis. Since these species are potential prey of tub
gurnard, these targets might have been overlooked during the stereo-microscopical analysis.
Although the advanced digestion of their remains in the fish stomach was apparent, these
species could still be detectable with the developed microarray assay.
As the feeding behaviour of fish might undergo spatio-temporal changes and the relative
frequency between different prey species would be quite variable, quantification would be
desirable; however, this is not possible at the current stage of INV-CHIP development.
Methods already used to quantify microbial populations [75] can be adapted to marine
organisms to quantitatively assess the diverse composition of natural populations. However,
the INV-CHIP can be used for regular monitoring of the gut content for ecosystem
functioning analysis and to investigate the common prey of commercially-relevant fishery
resources.
Developing a microarray requires a particularly demanding initial phase of information
gathering, in terms of sequence analysis of target and non-target species [76]. The critical
threshold is to obtain the most comprehensive database as possible, on which the successive
steps of probe design and in silico evaluation of oligos is based. Achieving such a sequence
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
23
database for the target invertebrate species of the present research was not a trivial task, since
only little information were available in terms of gene markers and molecular protocols. It
was not possible to conduct a preliminary in silico assessment of the variation of selected
genes at the between- and within-species levels due to few taxa and groups being well-
represented in the public databases at the time this study was performed. The lack of a proper
information background in terms of sequence variability at higher taxonomic levels (genus,
family or higher) also influenced the strategy of downstream probe design. However, DNA
barcoding is continuously increasing the inventory of animal biodiversity and, given the
promising performance of COI marker in the INV-CHIP, exploiting the BOLD sequence
information further probes could be successfully designed for targeting single species or
groups of species of high interest in future.[77][78]
Figure 4. Boxplots graphs summarizing the gut contents hybridization experiments. Sample codes given
as in Table 6.
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
24
Probe specificity is the first criterion to be considered during the in silico design and
selection of capture probes. Since the sequences used in this chapter represented only a small
subset of the gene diversity among individuals/populations of a given species, false negative
signals were expected either from non-inventoried sequence variants of target species, or
from unspecific binding with sequences of non-target species uncovered by the developed
microarray. Even if the non-target sequences dataset included as much of the gene diversity at
the species level of the taxa available in the public sequence databases, it cannot represent the
real and comprehensive species diversity of the invertebrate taxa. Therefore, non-target
DNAs from unknown and/or non-sequenced species can give false positive signals in the
hybridization of environmental samples. Moreover, single nucleotide polymorphisms can
occur within target species, due to natural biodiversity. In an attempt to reduce such
problems, redundant probes were designed for each species and no probes were removed
from the microarray even if their signal intensity was found to be very low.
A major concern in using DNA microarray for the detection and monitoring of
environmental samples is the misidentification of species due to cross-hybridization and non-
specific binding [50]. This might be caused by the high number of unknown organisms that
can be present in environmental samples. On the contrary, capture probes on the microarray
were designed only fora limited number of known and sequenced organisms; therefore, it is
impossible to detect the organisms not included as targets on the array [79]. In environmental
applications, DNA microarrays can also fail in identifying species if unknown haplotypes and
deep phylogeographic structure variants are present [80]. In this case, the development of a
“PhyloChip” microarray approach might help to overcome these empirical constraints [52].
The preparation of PCR amplicons as DNA targets required a considerable effort in terms
of setting up best conditions of DNA extraction, PCR amplification, and direct sequencing.
The taxonomic differences among selected target organisms (belonging to three taxa
Crustacea, Mollusca, and Polychaeta) must be considered in relation to the strategy to use
universal primers for multi-target amplification. Finally, the use of PCR amplicons as targets
in the microarray experiments was controversial, as the differential amplification of targets
could produce incorrect quantification of target and consequently introduce bias in the
hybridization results [78][81]. An accurate relationship between the amount of target and the
intensity of hybridization signal is obtained only using the isolated RNA for direct
hybridization on the microarray, avoiding the PCR amplification [82]. However, the RNA is
difficult to extract, subject to degradation, and with a rather low concentration in
environmental samples [82]. The RNA content per cell may vary under different
environmental conditions and different users could possibly isolate different amounts of
RNA. Therefore, the microarray-based quantification can be considered only as semi-
quantitative.
The strategy used in this chapter was to PCR-amplify and label the target DNA sequences
from the total genomic DNA template for subsequent hybridization. The suitability of using
multiplexed 16S and COI amplicons for hybridization on the INV-CHIP was verified by
assessing the yield of both gene markers on the agarose gel and on the assumption that both
markers were amplified with similar ratio. The same approach was also used to amplify both
genes from two-, three-, and mixed- templates, in the sensitivity experiments. The 16S
amplicons from multiple target species could be discriminated on the gel, as they did vary
between different species because of indels in their sequences (looped structure). Assuming
that the amplification would occur in the desired ratios, further assessments were carried out
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
25
at the level of the hybridization results, but the overall intensity of the signal was very low.
The reasons of such behavior could be related to competition among the primer pairs and
their differential efficiency to bind the template DNA. Another explanation could be due to
the high number of amplification cycles (40) used for labeling, which could produce a plateau
status in the PCR yield where amplification rate cannot be estimated. Nevertheless, all targets
used in the multiplex PCR gave positive hybridization signals on the INV-CHIP, suggesting
that all template DNA targets were amplified. For more quantitative results, a calibration
method is needed to estimate PCR efficiency, since the advantage of using universal PCR
primers for all targets should not be underestimated. Such a method would be very beneficial
for field applications since an efficient multiplex PCR is essential to amplify the target DNA
present in the environmental samples.
Data analysis and interpretation is one of the most critical methodological steps in the
interpretation of the hybridization results of microarrays [83]. The different types of
technology available pose new challenges to define standardized methods to process
microarray information. Unlike gene expression, where numerous algorithms exist for
statistical analysis, normalization, fold change analysis and biological interpretation, no
statistical tools have been specifically developed for the analysis of DNA microarray data.
Thus, the development of statistical algorithms for the analysis of quantitative data on the
abundance of the species in a mixture of environmental samples represents a challenge. A
typical analytical framework include data standardization, image acquisition and analysis,
normalization, exploratory data analysis, statistical significance inference, as well as various
considerations relevant to their implementation [56].
After the image acquisition, the data quality check can provide information on outlier
spots and identification of problematic slides. Data quality check is an essential part of the
data analysis, but it is still not easy to perform objectively or in an automated manner. Even
though techniques for the removal of outlier spots and to improve data quality are still under
development, some procedures seem to be more promising [84]. However, the usefulness of
many quality control measures is unsubstantiated and no specific quality control method has
been standardized yet [85].
The choice of the most appropriate statistical test that permit discrimination between low
and high abundant species against a background of non-targets is a challenging issue. The
detection of a meaningful response from the hybridization signals obtained in defined
conditions is often accomplished from the comparison of signal intensity between reference
and target spots, and by evaluating the significance of the deviation of signal intensity in
relation to a fixed threshold. Methods of data analysis relying on the intensity threshold
appear to be very limited from a statistical point of view. Therefore, a Student's t-test, with
appropriate correction for unequal variances, was adopted to compare local background with
negative controls. A one-way ANOVA analysis allowed testing the significance of
hybridization signals versus controls. In most cases, fluorescence signals showed very large
variance, which commonly cause problems in the application of parametric data analysis
techniques. ANOVA relies on the assumption of heteroscedasticity [86]. Data transformations
are commonly adopted to reduce the effects of heterogeneous variances in the comparative
analysis of groups of samples. Though it is a common practice to perform statistical analysis
on a log-transformed data, variances may remain heterogeneous even after transformation. In
these cases some authors [67] suggest to perform analyses with untransformed data. When
departures from assumptions of ANOVA occur, several authors prefer to switch to non-
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
26
parametric methods, such as Kruskal-Wallis test, even if variance heterogeneity still can
weaken consistency of the results [87]. In this chapter, a comparison of tests carried out both
with transformed and untransformed data showed that in the former case a dramatic increase
in false positives occurred (data not shown). Therefore, untransformed data were analysed.
The identification of target hybridization signals higher than those of the controls was
performed using a one-way ANOVA with a customized pairwise one-tailed t-tests approach.
Multiple post-hoc tests were performed following planned comparisons procedure between
controls and probes (like the Dunnett's procedure), with p-values correction made by
Benjamini-Hochberg's FDR (False Discovery Rate) [68].
The adoption of this pairwise procedure was decided after an accurate comparison of
results of several tests most commonly recommended in literature for data with unequal
variances and unequal sample sizes. The detection of false positive and negative signals has
revealed to be one of the most tedious problems to solve. In this chapter, the t-test corrected
with FDR performed very satisfactorily in terms of avoiding Type I errors, maintaining a
good power of the test (low rate of false negatives) even in presence of an unbalanced design
or small sample size.
A thoughtful experimental design is the most important step at the beginning of a
microarray experiment. A well-planned experiment can maximize the efficiency of
subsequent data analysis. Experimental design should include an adequate number of
replicates at different levels to separate effects of different sources of variation, which may be
technical (e.g. dye, batch) or biological (e.g. genotype, sex, time) [55]. Probe level variation
carries important information useful to control technical sources of variation, outlier removal,
and subsequent statistical analyses [88]. In some cases, the technological platform can permit
increases in the technical replication and model the statistical properties of the data [89], thus
improving the quality of the analysis. The control of biological variation is also challenging
not only for analytical procedures, but also for costs associated with sampling of additional
individuals [55].
The adoption of a balanced experimental design can help to avoid difficulties about the
choice of an appropriate test, reducing the probability of Type I and Type II errors. Although
there is no consensus about what is the best procedure to define the optimal sample size, there
is consensus that power analyses should be performed applying the newer methods
specifically implemented for microarray data, and that more replicates generally provide
greater power [55].
Challenges in data analysis are often linked to the development of new technologies.
Microarray analysis is evolving rapidly, as new and more complex analyses appear, making it
easy for the researcher to get lost in endless new methods and software. Collaborations
between biologists, statisticians, and mathematicians is advisable for proper planning and
analyses [56]. Commercial and open-source software for performing data analyses are now
widely available. The development of dedicated R packages [69] and increasing availability
of libraries for the analysis of microarray can help researchers apply "canned" procedures as
well as to customize data analysis tools according to the particular needs of the data structure.
Although DNA chip technology was applied for the first time nearly 30 years ago [30] to
discriminate six DNA genotypes and nine common allelic mutations in one single
experiment, this technique has helped to circumvent much of the problems associated with the
conventional techniques used in taxonomy. The past decade experienced great advances and
changes in molecular methods for species identification, spanning from approaches allowing
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
27
the identification of one single specimen at once, to high-throughput techniques allowing the
identification of hundreds or thousands of specimens at a time. The DNA barcoding
approach, first introduced in 2003 [15], allows species identification by analysing single short
sequences of the mitochondrial COI and compare it to all information available in public
libraries [90][91]. The high-throughput next generation sequencing (NGS) technologies firstly
implemented in 2005 [92][93] can perform massive, parallel analyses of a vast amount of
specimens and species, even from fragmented DNA. NGS has revolutionized genomics
research by providing large amounts of information quickly and efficiently [94]. NGS have
recently been applied in species identification and ecological studies, since NGS platforms
allow the direct analyses of DNA sequences from environmental samples [95][96].
Complete genome re-sequencing using NGS provides greater depth in terms of sample
sequence coverage and helps in identifying unknown diversity in ecological communities
[97][98]. Compared to NGS, the disadvantage of using microarrays in environmental and
ecological studies is due to the identification of samples restricted to those species for which
probes are captured on the array, and are limited in recognizing unknown, or novel species
[99][100]. The development of microarray for new species identification and sample
processing requires an intense procedure of benchmarking and optimization that outperform
results.
Using microarray as a validation tool of NGS data is quite common when amplicon
information from NGS is used for probe designing in microarray experiments for species
identification [101]. NGS is the most preferred technology as data from one single experiment
can provide in-depth analysis of the species diversity [102][103]. In some applications, DNA
microarrays are sought after over NGS for routine identification of species in which sequence
information is already known and specific probes are designed for the microarray. Despite
drawbacks, microarrays have already been developed for species identification mainly in
ecological and biodiversity assessment studies of microbial communities and algal blooms
[34][43][36]. Overcoming the above mentioned problems would further increase the
applicability of microarrays as an identification tool for identifying any organism down to
species level [15][80].
ACKNOWLEDGMENTS
The research leading to these results has received funding from the European
Community’s Sixth Framework Programme (http://ec.europa.eu/research/fp6/index_en.cfm),
as Specific Targeted Research Project (STREP) funded by the European Commission under
the contract no. 505491: “Fish & Chips”: Towards DNA chip technology as a standard
analytical tool for the identification of marine organisms in biodiversity and ecosystem
research (www.fish-and-chips.uni-bremen.de).
The authors thank the whole Fish & Chips consortium for many insightful discussions. S.
Roll and F. Meyerjürgens produced the microarrays. Hannes Weber and Mangred Noelte
developed CDIS and designed oligonucleotide probes for the INV-CHIP. Antonios Magoulas
and Christos Arvanitidis, HCMR Crete helped taxonomical identification of Polychaete
species. Venugopal Moleyur, College of Fisheries, Mangalore, India helped in INV CHIP
optimization experiments.
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
28
REFERENCES
[1] Vermeulen, N. From darwin to the census of marine life: marine biology as big science.
PloS One, 2013, 8, e54284.
[2] Bianchi, CN; Morri, C. Marine biodiversity of the Mediterranean Sea: situation,
problems and prospects for future research. Mar Poll Bull, 2000, 40, 367–376.
[3] Bortolus, A. Error cascades in the biological sciences: the unwanted consequences of
using bad taxonomy in ecology. J Hum Environ, 2008, 37, 114–118.
[4] Daan, N. Data base report of the stomach sampling project 1981. ICES Cooperative
Research Report No 164. International Council for the Exploration of the Sea, 1989.
[5] Hislop, J; Bromley, PJ; Daan, N; Gislason, H; Heessen, HJL; Robb, AP;et al. Database
report of the Stomach Sampling Project, 1991. ICES Cooperative Research Report No.
219. International Council for the Exploration of the Sea, 1997.
[6] Hyslop, EJ. Stomach contents analysis-a review of methods and their application. J
Fish Biol, 1980, 17, 411–429.
[7] Gorokhova, E; Fagerberg, T; Hansson, S. Predation by herring (Clupea harengus) and
sprat (Sprattus sprattus) on Cercopagis pengoi in a western Baltic Sea bay. ICES J Mar
Sci, 2004, 61, 959–965.
[8] Espinoza, M; Wehrtmann, IS. Stomach content analyses of the threadfin anglerfish
Lophiodes spilurus (Lophiiformes: Lophiidae) associated with deepwater shrimp
fisheries from the central Pacific of Costa Rica. Revista de Biología Tropical, 2008, 56,
1959–1970.
[9] Cocheret, de; la Morinière, E; Pollux, B; Nagelkerken, I; Hemminga, M; Huiskes, A;
van der Velde, G. Ontogenetic dietary changes of coral reef fishes in the mangrove-
seagrass-reef continuum: stable isotopes and gut-content analysis. Mar Ecol Prog Ser,
2003, 246, 279–289.
[10] Rosel, P; Kocher, T. DNA-based identification of larval cod in stomach contents of
predatory fishes. J Exp Mar Biol Ecol, 2002, 267, 75–88.
[11] Baker, R,; Buckland, A; Sheaves, M. Fish gut content analysis: robust measures of diet
composition. Fish Fish, 2013, doi, 10.1111/faf.12026.
[12] Rindorf, A; Lewy, P. Bias in estimating food consumption of fish from stomach-content
analysis. Can J Fish Aquat Sci, 2004, 61, 2487–2498.
[13] Hynes, HBN. The food of fresh-water sticklebacks (Gasterosteus aculeatus and
Pygosteus pungitius), with a review of methods used in studies of the food of fishes. J
Anim Ecol, 1950, 19, 36–58.
[14] Bowen, S. Quantitative description of the diet. In: Nielsen LA, Johnson DL (eds).
Fisheries techniques. American Fisheries Society, Bethesda, Maryland, 1983, 325–336.
[15] Hebert, PDN; Cywinska, A; Ball, SL; deWaard, JR. Biological identifications through
DNA barcodes. Proc Biol Sci R Soc, 2003, 270, 313–321.
[16] Hebert, PDN; Stoeckle, MY; Zemlak, TS; Francis, CM. Identification of birds through
DNA barcodes. PLoS Biology, 2004, 2, e312.
[17] Hebert, PDN; Gregory, TR. The promise of DNA barcoding for taxonomy. Syst Biol,
2005, 54, 852–859.
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
29
[18] Kochzius, M; Kappel, K; Dobitz, L; Silkenbeumer, N; Nolte, M; Weber, H;et al. The
“Fish & Chips” project: microarrays as a tool for the identification of marine organisms
in biodiversity and ecosystem research. OCEANS 2007 - Europe, IEEE,2007,1–4.
[19] Féral, J. How useful are the genetic markers in attempts to understand and manage
marine biodiversity? J Exp Mar Biol Ecol, 2002, 268, 121–145.
[20] Ward, RD; Hanner, R; Hebert, PDN. The campaign to DNA barcode all fishes, FISH-
BOL. J Fish Biol, 2009, 74, 329–356.
[21] Jérôme, M; Martinsohn, JT; Ortega, D; Carreau, P; Verrez-Bagnis, V; Mouchel, O.
Toward fish and seafood traceability: anchovy species determination in fish products by
molecular markers and support through a public domain database. J Agr Food Chem,
2008, 56, 3460–3469.
[22] Teletchea, F. Molecular identification methods of fish species: reassessment and
possible applications. Rev Fish Biol Fisheries, 2009, 19, 265–293.
[23] Espiñeira, M; González-Lavín, N; Vieites, JM; Santaclara, FJ. Development of a
method for the genetic identification of flatfish species on the basis of mitochondrial
DNA sequences. J Agr Food Chem, 2008, 56, 8954–8961.
[24] Costa, FO,; Carvalho, GR. The Barcode of Life Initiative: synopsis and prospective
societal impacts of DNA barcoding of Fish. Genomics Soc Policy, 2007, 3, 29-40.
[25] Rocha-Olivares, A. Multiplex haplotype-specific PCR: a new approach for species
identification of the early life stages of rockfishes of the species-rich genus Sebastes
Cuvier. J Exp Mar Biol Ecol, 1998, 231, 279–290.
[26] Sweijd, NA; Bowie, RCK; Evans, BS; Lopata, AL. Molecular genetics and the
management and conservation of marine organisms. Hydrobiologia, 2000, 420,
153–164.
[27] Bucklin, A; Guarnieri, M; Hill, RS; Bentley, AM; Kaartvedt, S. Taxonomic and
systematic assessment of planktonic copepods using mitochondrial COI sequence
variation and competitive, species-specific PCR. Hydrobiologia, 1999, 401, 239–254.
[28] Bell, JL; Grassle, JP. A DNA probe for identification of larvae of the commercial
surfclam (Spisula solidissima). Mol Mar Biol Biotechnol, 1998, 7, 127–137.
[29] Kochzius, M; Nölte, M; Weber, H; Silkenbeumer, N; Hjörleifsdottir, S; Hreggvidsson,
GO;et al. DNA microarrays for identifying fishes. Mar Biotechnol, 2008, 10, 207–217.
[30] Saiki, RK. Genetic analysis of amplified DNA with immobilized sequence-specific
oligonucleotide probes. Proc Natl Acad Sci USA, 1989, 86, 6230–6234.
[31] Southern, EM. Detection of specific sequences among DNA fragments separated by gel
electrophoresis. J Mol Biol, 1975, 98, 503–517.
[32] Lashkari, DA; DeRisi, JL; McCusker, JH; Namath, AF; Gentile, C; Hwang, SY;et al.
Yeast microarrays for genome wide parallel genetic and gene expression analysis. Proc
Natl Acad Sci, USA, 1997, 94, 13057–13062.
[33] Call, DR; Borucki, MK; Loge, FJ. Detection of bacterial pathogens in environmental
samples using DNA microarrays. J Microbiol Methods, 2003, 53, 235–243.
[34] Franke-Whittle, IH,; Klammer, SH; Insam, H. Design and application of an
oligonucleotide microarray for the investigation of compost microbial communities. J
Microbiol Methods, 2005, 62, 37–56.
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
30
[35] Loy, A; Bodrossy, L. Highly parallel microbial diagnostics using oligonucleotide
microarrays. Clinica Chimica Acta, 2006, 363, 106–119.
[36] Peplies, J; Lau, SCK; Pernthaler, J; Amann, R; Glöckner, FO. Application and
validation of DNA microarrays for the 16S rRNA-based analysis of marine
bacterioplankton. Environ Microbiol, 2004, 6, 638–645.
[37] Panicker, G; Vickery, MCL; Bej, AK. Multiplex PCR detection of clinical and
environmental strains of Vibrio vulnificus in shellfish. Can J Microbiol, 2004, 50,
911–922.
[38] González, SF; Krug, MJ; Nielsen, ME; Santos, Y; Call, DR. Simultaneous detection of
marine fish pathogens by using multiplex PCR and a DNA microarray. J Clin
Microbiol, 2004, 42, 1414–1419.
[39] Yoon, HK; Jeong, D; Chung, IH; Jung, JW; Oh, MJ; Kim, S. Rapid species
identification of elasmobranch fish (skates and rays) using oligonucleotide microarray.
Biochip, 2009, 3, 87–96.
[40] Kochzius, M; Seidel, C; Antoniou, A; Botla, SK; Campo, D; Cariani, A;et
al.Identifying fishes through DNA barcodes and microarrays. PloS One, 2010, 5,
e12620.
[41] Metfies, K; Medlin, LK. DNA microchips for phytoplankton: The fluorescent wave of
the future. Nova Hedwigia, 2004, 79, 321–327.
[42] Metfies, K; Berzano, M; Mayer, C; Roosken, P; Gualerzi, C; Medlin, L;et al. An
optimized protocol for the identification of diatoms, flagellated algae and pathogenic
protozoa with phylochips. Molecular Ecol Notes, 2007, 7, 925–936.
[43] Gescher, C; Metfies, K; Medlin, LK. The ALEX CHIP—Development of a DNA chip
for identification and monitoring of Alexandrium. Harm Algae, 2008, 7, 485–494.
[44] Galluzzi, L; Cegna, A; Casabianca, S; Penna, A; Saunders, N; Magnani, M.
Development of an oligonucleotide microarray for the detection and monitoring of
marine dinoflagellates. J Microbiol Methods, 2011, 84, 234–242.
[45] McCoy, GR; Touzet, N; Fleming, GT; Raine, R. An evaluation of the applicability of
microarrays for monitoring toxic algae in Irish coastal waters. Environ Sci Poll Res Int,
2012, doi 10.1007/s11356-012-1294-1.
[46] Kegel, J; Del Amo, Y; Costes, L; Medlin, L. Testing a microarray to detect and monitor
toxic microalgae in Arcachon Bay in France. Microarrays, 2013, 2, 1–23.
[47] Pauly, D. Fishing down marine food webs. Science, 1998, 279, 860–863.
[48] Goffredi, SK; Jones, WJ; Scholin, CA; Marin, R; Vrijenhoek, RC. Molecular detection
of marine invertebrate larvae. Mar Biotechnol, 2006, 8, 149–160.
[49] Kothapalli, R; Yoder, S; Mane, S; Loughran, T. Microarray results: how accurate are
they? BMC Bioinformatics, 2002, 3, 22.
[50] Zhou, J; Thompson, DK. Challenges in applying microarrays to environmental studies.
Curr Opin Biotechnol, 2002, 13, 204–207.
[51] Gescher, C; Metfies, K; Frickenhaus, S; Knefelkamp, B; Wiltshire, KH; Medlin, LK.
Feasibility of assessing the community composition of prasinophytes at the Helgoland
Roads sampling site with a DNA microarray. Appl Environ Microbiol, 2008, 74, 5305–
5316.
[52] Metfies, K; Medlin, LK. Feasibility of transferring fluorescent in situ hybridization
probes to an 18S rRNA gene phylochip and mapping of signal intensities. Appl Environ
Microbiol, 2008, 74, 2814–2821.
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
31
[53] Kim, S; Eo, H-S; Koo, H; Choi, J-K; Kim, W. DNA barcode-based molecular
identification system for fish species. Mol Cells, 2010, 30, 507–512.
[54] Peplies, J; Glöckner, FO; Amann, R. Optimization strategies for DNA microarray-
based detection of bacteria with 16S rRNA-targeting oligonucleotide probes. Appl
Environ Microbiol, 2003, 69, 1397–1407.
[55] Kerr, MK. Design considerations for efficient and effective microarray studies.
Biometrics, 2003, 59, 822–828.
[56] Leung, YF; Cavalieri, D. Fundamentals of cDNA microarray data analysis. Trends
Genet, 2003, 19, 649–659.
[57] Deagle, BE; Tollit, DJ; Jarman, SN; Hindell, MA; Trites, AW; Gales, NJ. Molecular
scatology as a tool to study diet: analysis of prey DNA in scats from captive Steller sea
lions. Mol Ecol, 2005, 14, 1831–1842.
[58] Folmer, O; Black, M; Hoeh, W; Lutz, R; Vrijenhoek, R. DNA primers for amplification
of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates.
Mol Mar Biol Biotechnol, 1994, 3, 294–299.
[59] Palumbi, S; Romano, S; Mcmillan, WO; Grabowski, G. The simple fool’s guide to
PCR, 2002, 1–45. http://palumbi.stanford.edu/SimpleFoolsMaster.pdf.
[60] Tamura, K; Peterson, D; Peterson, N; Stecher, G; Nei, M; Kumar, S. MEGA5:
molecular evolutionary genetics analysis using maximum likelihood, evolutionary
distance, and maximum parsimony methods. Mol Biol Evol, 2011, 28, 2731–2739.
[61] Altschul, SF; Gish, W; Miller, W; Myers, EW; Lipman, DJ. Basic local alignment
search tool. J Mol Biol, 1990, 215, 403–410.
[62] Nolte, M. Optimization of oligonucleotide sets for DNA microarrays. PhD Thesis
University of Bremen, Germany, 2002.
[63] Hofacker, IL; Fontana, W; Stadler, PF; Bonhoeffer, LS; Tacker, M; Schuster, P. Fast
folding and comparison of RNA secondary structures. Monatshefte Für
Chemie/Chemical Monthly, 1994, 125, 167–188.
[64] Stagioni, M; Montanini, S; Vallisneri, M. Feeding of tub gurnard Chelidonichthys
lucerna (Scorpaeniformes: Triglidae) in the north-east Mediterranean. J Mar Biol Assoc
UK, 2011, 92, 605–612.
[65] Welch, BL. On the comparison of several mean values: An alternative approach.
Biometrika, 1951, 38, 330–336.
[66] Snedecor, GW;Cochran, WG. Statistical Methods. 8th Editio. Iowa State University
Press; 1989.
[67] Underwood, A. Techniques of analysis of variance in experimental marine biology and
ecology. Oceanogr Mar Biol, 1981, 16.
[68] Benjamini, Y; Hochberg, Y. Controlling the false discovery rate: a practical and
powerful approach to multiple testing. J R Stat Soc B, 1995, 57, 289–300.
[69] R Development Core Team. {R: A language and environment for statistical
computing}. Vienna, Austria: R Foundation for Statistical Computing; 2009.
[70] Bucklin, A; Frost, B; Bradford-Grieve, J; Allen, L; Copley, N. Molecular systematic
and phylogenetic assessment of 34 calanoid copepod species of the Calanidae and
Clausocalanidae. Mar Biol, 2003, 142, 333–343.
[71] Brown, B; Emberson, RM; Paterson, AM. Mitochondrial COI and II provide useful
markers for Wiseana (Lepidoptera: Hepialidae) species identification. Bull EntomolRes,
1999, 89, 287–293.
Srujana Chitipothu, Alessia Cariani, Fabio Bertasi et al.
32
[72] Janzen, DH; Hajibabaei, M; Burns, JM; Hallwachs, W; Remigio, E; Hebert, PDN.
Wedding biodiversity inventory of a large and complex Lepidoptera fauna with DNA
barcoding. Philos Trans R Soc Lond Biol Sci, 2005, 360, 1835–1845.
[73] Shouche, YS; Patole, MS. Sequence analysis of mitochondrial 16S ribosomal RNA
gene fragment from seven mosquito species. J Biosci, 2000, 25, 361–366.
[74] Barrett, RD; Hebert, PD. Identifying spiders through DNA barcodes. Can J Zool, 2005,
83, 481–491.
[75] Palmer, C; Bik, EM; Eisen, MB; Eckburg, PB; Sana, TR; Wolber, PK;et al. Rapid
quantitative profiling of complex microbial populations. Nucl Acids Res, 2006, 34, e5.
[76] Schena, M; Shalon, D; Davis, RW; Brown, PO. Quantitative monitoring of gene
expression patterns with a complementary DNA microarray. Science, 1995, 270,
467–470.
[77] Ratnasingham, S; Hebert, PDN. bold: The Barcode of Life Data System
(http://www.barcodinglife.org). Mol Ecol Notes, 2007, 7, 355–364.
[78] Kanagawa, T. Bias and artifacts in multitemplate polymerase chain reactions (PCR). J
Biosci Bioeng, 2003, 96, 317–323.
[79] Gentry, TJ; Wickham, GS; Schadt, CW; He, Z; Zhou, J. Microarray applications in
microbial ecology research. Microb Ecol, 2006, 52, 159–175.
[80] Hajibabaei, M; Singer, GAC; Hebert, PDN; Hickey, DA. DNA barcoding: how it
complements taxonomy, molecular phylogenetics and population genetics. Trends
Genet, 2007, 23, 167–172.
[81] Medlin, LK; Metfies, K; Mehl, H; Wiltshire, K; Valentin, K. Picoeukaryotic plankton
diversity at the Helgoland time series site as assessed by three molecular methods.
Microb Ecol, 2006, 52, 53–71.
[82] Peplies, J; Lachmund, C; Glöckner, FO; Manz, W. A DNA microarray platform based
on direct detection of rRNA for characterization of freshwater sediment-related
prokaryotic communities. Appl Environ Microbiol, 2006, 72, 4829–4838.
[83] Blohm, DH; Guiseppi-Elie, A. New developments in microarray technology. Curr Opin
Biotechnol, 2001, 12, 41–47.
[84] Kauffmann, A; Huber, W. Microarray data quality control improves the detection of
differentially expressed genes. Genomics, 2010, 95, 138–142.
[85] Allison, DB; Cui, X; Page, GP; Sabripour, M. Microarray data analysis: from disarray
to consolidation and consensus. Nat Rev Genet, 2006, 7, 55–65.
[86] Sokal, RR;Rohlf, FJ. Biometry. 3rd Editio. Newyork Freeman, 1995.
[87] Hollander, M; Wolfe, DA. Nonparametric statistical methods. Wiley, 1973.
[88] Ayroles, JF; Gibson, G. Analysis of variance of microarray data. Methods Enzymol,
2006, 411, 214–233.
[89] Lin, SM; Du, P; Huber, W; Kibbe, W; a. Model-based variance-stabilizing
transformation for Illumina microarray data. Nucl Acids Res, 2008, 36, e11.
[90] Costa, FO; Landi, M; Martins, R; Costa, MH; Costa, ME; Carneiro, M;et al.A ranking
system for reference libraries of DNA barcodes: application to marine fish species from
Portugal. PloS One, 2012, 7, e35858.
[91] Bergsten, J; Bilton, DT; Fujisawa, T; Elliott, M; Monaghan, MT; Balke, M;et al. The
effect of geographical scale of sampling on DNA barcoding. Syst Biol, 2012, 61,
851–869.
Invertebrate DNA Chip: Opportunities and Challenges in the Development …
33
[92] Margulies, M; Egholm, M; Altman, WE; Attiya, S; Bader, JS; Bemben, La;et al.
Genome sequencing in microfabricated high-density picolitre reactors. Nature, 2005,
437, 376–380.
[93] Shendure, J; Porreca, GJ; Reppas, NB; Lin, X; McCutcheon, JP; Rosenbaum, AM;et al.
Accurate multiplex polony sequencing of an evolved bacterial genome. Science, 2005,
309, 1728–1732.
[94] Morozova, O; Marra, MA. Applications of next-generation sequencing technologies in
functional genomics. Genomics, 2008, 92, 255–264.
[95] Ekblom, R; Galindo, J. Applications of next generation sequencing in molecular
ecology of non-model organisms. Heredity, 2011, 107, 1–15.
[96] Sogin, ML; Morrison, HG; Huber, JA; Mark Welch, D; Huse, SM; Neal, PR;et al.
Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc Natl
Acad Sci, USA 2006, 103, 12115–12120.
[97] Kermarrec, L; Franc, A; Rimet, F; Chaumeil, P; Humbert, JF; Bouchez, A. Next-
generation sequencing to inventory taxonomic diversity in eukaryotic communities: a
test for freshwater diatoms. Mol Ecol Resources, 2013, 13, 607–619.
[98] Bokulich, NA; Joseph, CML; Allen, G; Benson, AK; Mills, DA. Next-generation
sequencing reveals significant bacterial diversity of botrytized wine. PloS One, 2012, 7,
e36357.
[99] Gilad, Y; Rifkin, SA; Bertone, P; Gerstein, M; White, KP. Multi-species microarrays
reveal the effect of sequence divergence on gene expression profiles. Genome Res,
2005, 15, 674–680.
[100] Kammenga, JE; Herman, MA; Ouborg, NJ; Johnson, L; Breitling, R. Microarray
challenges in ecology. Trends Ecol Evol, 2007, 22, 273–279.
[101] Roh, SW; Abell, GCJ; Kim, KH; Nam, YD; Bae, JW. Comparing microarrays and next-
generation sequencing technologies for microbial ecology research. Trends Biotechnol,
2010, 28, 291–299.
[102] Marinković, M; de Leeuw, WC; de Jong, M; Kraak, MHS; Admiraal, W; Breit, TM;et
al. Combining next-generation sequencing and microarray technology into a
transcriptomics approach for the non-model organism Chironomus riparius. PloS One,
2012, 7, e48096.
[103] Park, JY; Lee, SY;, An, CM; Kang, JH; Kim, JH; Chai, JC;et al. Comparative study
between Next Generation Sequencing Technique and identification of microarray for
Species Identification within blended food products. BioChip J, 2012, 6, 354–361.