GenomeCarver: harvesting genetic parts from genomes to support biological design automation

Abstract

Concept The advance of genome sequencing and annotation has provided a " gold mine " of genetic parts, which all synthetic biologists wish to include in their toolbox of parts with which to build synthetic biological systems. The currently available computer assisted design systems (CADs) focus heavily, if not exclusively, on composing biological systems using genetic parts [3][7][9], however, how a user obtains parts in the first place remains an open question. To make matters worse, there are a few dozen part standards being proposed and used in the synthetic biology community (104 RFCs on part standard as of today). Even though one can extract a few parts from the genome manually, there is no software to ensure the standard compatibility of parts, and it is also very dicult to scale up the design of parts. With these problems in mind, we present GenomeCarver, a computational tool for the harvesting and packaging of biological parts from model genomes. GenomeCarver interfaces with various genomes, identifies regions of interest according to user specification (e.g., promoters, open reading frames and terminators) and extraction rules (e.g., a promoter is defined as 500bp upstream of the ATG start codon or last gene boundary, which comes shorter), extracts corresponding DNA sequences from the genome feature files (GFFs), checks the sequence's compatibility with the selected standard (e.g., whether the given sequence includes the forbidden restriction sites of certain parts standards), and finally outputs optimized primer sequences to amplify the parts from genomic DNA, adding necessary flanking sequences to standardize the parts. Through its compatibility with multiple genomes and multiple parts standards, GenomeCarver bridges the fields of systems biology and synthetic biology, and greatly enriches synthetic biologists' design toolbox. It complements many parts-based design tools which currently exist by supporting the Synthetic Biology Open Language standard [6]. Implementation GenomeCarver can be accessed as an application built on ⇤ These authors contributed equally to this work.

Full text

GenomeCarver: harvesting genetic parts from genomes to
support biological design automation
Emily Scher⇤
School of Informatics
University of Edinburgh
Edinburgh EH9 3JR, UK
Yisha Luo⇤
School of Biological Sciences
University of Edinburgh
Edinburgh EH9 3JR, UK
Aaron Berliner
Autodesk Research
San Francisco, California
94111, USA
Jacqueline Quinn
Autodesk Research
San Francisco, California
94111, USA
Carlos Olguin
Autodesk Research
San Francisco, California
94111, USA
Dr. Yizhi Cai†
School of Biological Sciences
University of Edinburgh
Edinburgh EH9 3JR, UK
ABSTRACT
Concept
The advance of genome sequencing and annotation has pro-
vided a “gold mine” of genetic parts, which all synthetic biol-
ogists wish to include in their toolbox of parts with which to
build synthetic biological systems. The currently available
computer assisted design systems (CADs) focus heavily, if
not exclusively, on composing biological systems using ge-
netic parts [3][7][9], however, how a user obtains parts in
the first place remains an open question. To make matters
worse, there are a few dozen part standards being proposed
and used in the synthetic biology community (104 RFCs on
part standard as of today). Even though one can extract a
few parts from the genome manually, there is no software
to ensure the standard compatibility of parts, and it is also
very difficult to scale up the design of parts.
With these problems in mind, we present GenomeCarver, a
computational tool for the harvesting and packaging of bio-
logical parts from model genomes. GenomeCarver interfaces
with various genomes, identifies regions of interest according
to user specification (e.g., promoters, open reading frames
and terminators) and extraction rules (e.g., a promoter is
defined as 500bp upstream of the ATG start codon or last
gene boundary, which comes shorter), extracts correspond-
ing DNA sequences from the genome feature files (GFFs),
checks the sequence’s compatibility with the selected stan-
dard (e.g., whether the given sequence includes the forbid-
den restriction sites of certain parts standards), and finally
outputs optimized primer sequences to amplify the parts
from genomic DNA, adding necessary flanking sequences to
standardize the parts.
Through its compatibility with multiple genomes and mul-
tiple parts standards, GenomeCarver bridges the fields of
systems biology and synthetic biology, and greatly enriches
synthetic biologists’ design toolbox. It complements many
parts-based design tools which currently exist by supporting
the Synthetic Biology Open Language standard [6].
Implementation
GenomeCarver can be accessed as an application built on
⇤These authors contributed equally to this work.
†Corresponding author. E-mail:yizhi.cai@ed.ac.uk
Autodesk’s Project Cyborg (http://autodeskresearch.com/
projects/cyborg). Project Cyborg is a cloud-based plat-
form for computational tools in the life sciences and pro-
grammable matter space, supporting design and engineering
across domains and scales. Cyborg enables elastic comput-
ing through a node framework that natively provides sup-
port for simulation, optimization, and visualization. Being
built on Cyborg, GenomeCarver is comprised of nodes for
each step of the workflow connected to form a cohesive user
experience that guides the user through the tool.
GenomeCarver currently supports three model organisms:
yeast Saccharomyces cerevisiae, bacterial Escherichia coli,
and plant Arabidopsis thaliana. However, GenomeCarver
is flexible enough to be extended to interface a variety of
organisms, which we plan to do in the near future. Simi-
larly, GenomeCarver currently supports a finite number of
mainstream parts standards such as the BioBrick 1.0[1] and
yeast Golden Gate standards[2], but new standards could
easily be incorporated. In a future implementation, we even
plan to allow users to import their own, custom standards.
While it’s interfacing with multiple genomes and standards
has made GenomeCarver flexible, it’s being built on top of
Cyborg further’s the tool’s flexibility, as GenomeCarver will
be able to be used in conjunction with the other tools cur-
rently being developed on the same platform.
Figure 1 shows the application’s workflow. First, a user
chooses a genome, a category and the loci of the part. For
instance, a user may choose the promoter of Gal loci from
Saccharomyces cerevisiae. Optionally, the user can then de-
fine the preferred promoter and terminator lengths, or spec-
ify that they would like gene boundaries to be ignored. The
default maximum promoter length is 500 base pairs, and
the default maximum terminator length is 200 base pairs.
If the user does not specify that gene boundaries should
be ignored, then GenomeCarver will identify a gene’s pro-
moter as the upstream (5’ to 3’) sequence of a maximum
length of 500 (or the specified maximum length) which does
not overlap another gene. It will identify the terminator as
the following sequence of a maximum length of 200 bases
which does not overlap another gene. GenomeCarver then
returns the specified sequence(s). The user can then as-
sign the sequence(s) to a standard using another drop down
menu. Once selected, the sequence will then be checked
16
IWBDA 2014, June 11–12, 2014, Boston, Massachusetts, USA.
Copyright is held by the owner/author(s). Publication rights licensed to BDAC.
IWBDA 2014, June 11-12, 2014, Boston, Massachusetts, USA. Copyright is held by the owner/author(s).
BDAC acknowledges that this contribution was authored or co-authored by a contractor or affiliate of the
U.S. Government. As such, the Government retains a nonexclusive, royalty-free right to publish or
reproduce this article, or to allow others to do so, for Government purposes only.

Figure 1: The workflow of GenomeCarver
for restriction sites, returning a warning if an incompatible
restriction site is found. The part will then be packaged
by adding the appropriate prefix and suffix. GenomeCarver
then allows the packaged parts to be exported in CSV and
SBOL formats [6].
Experimental verification
GenomeCarver has been used extensively in several labs
in the USA, UK and China to systematically design thou-
sands of yeast parts of each category conforming to the yeast
Golden Gate standard. We used the designed primers to am-
plify parts from genomic DNA in a high throughput fashion,
cloning the parts onto Topo vector backbone, and sequence
verifying them all (data not shown in the abstract). We
also demonstrated high efficient assembly of various genetic
switches using these parts and standard Golden Gate reac-
tion, and transformed these assembled switches into yeast
for functional assays. Most recently, GenomeCarver has
been used to design all the 6000 yeast promoters and 6000
yeast terminators, which demonstrates that we can scale up
the design automation easily.
Fut u r e p l a n
In the next version of GenomeCarver, we are planing to
include additional genomes, such as mammalian ones, as
well as to support user-customized standards. Batch de-
sign functionality will also be developed to support large
projects, such as BioFab (http://biofab.synberc.org/) type
projects for various genomes. We will also develop better
primer design strategies [8] to maximize the parts amplifi-
cation success rate. We are also planning to support codon
optimization for gene parts, so that a user can carve out
a gene from one species and codon optimize it for another
species, and GenomeCarve will output oligonucleotides for
de novo DNA synthesis. Finally, a better integration with
existing parts-based design tools will be needed for a better
user design experience.
Conclusion
GenomeCarver has been built to fill a gap left by existing
Synthetic Biology computational tools. It allows users to
extract parts directly from genomes, and to package them
into standardized formats for parts synthesis. We have used
this tool to design over 12,000 parts, and constructed and
verified several hundred of them. This tool, along with the
parts repository we created using it, will be a useful and
important addition to the synthetic biology community.
Acknowledgement
ES is supported by an Autodesk research fellowship. The
project is funded by an Edinburgh Chancellor’s Fellowship
to YC. We thank Drs. Jef D. Boeke (New York Univer-
sity, USA) and Junbiao Dai (Tsinghua University, China)
for helpful discussions to initiate the project.
1. REFERENCES
[1] Technical report, August 2005.
[2] Technical report, Johns Hopkins University, October
2012.
[3] Yizhi Cai, Brian Hartnett, Claes Gustafsson, and Jean
Peccoud. A syntactic model to design and verify
synthetic genetic constructs derived from standard
biological parts. Bioinformatics,2007.
[4] Carola Engler, Ramona Gruetzner, Romy Kandzia, and
Sylvestre Marillonnet. Golden gate shu✏ing: A one-pot
dna shu✏ing method based on type iis restriction
enzymes. PLoS ONE,4(5):e5553,052009.
[5] Carola Engler, Romy Kandzia, and Sylvestre
Marillonnet. A one pot, one step, precision cloning
method with high throughput capability. PLoS ONE,
3(11):e3647, 11 2008.
[6] Michal Galdzicki, Cesar Rodriguez, Deepak Chandran,
Herbert M. Sauro, and John H. Gennari. Standard
biological parts knowledgebase. PLoS ONE,
6(2):e17005, 02 2011.
[7] Nathan J. Hillson, Rafael D. Rosengarten, and Jay D.
Keasling. j5 dna assembly design automation software.
ACS Synthetic Biology,1(1):14–21,2012.
[8] Andreas Untergasser, Ioana Cutcutache, Triinu
Koressaar, Jian Ye, Brant C. Faircloth, Maido Remm,
and Steven G. Rozen. Primer3ˆa ˘
Aˇ
Tnew capabilities and
interfaces. Nucleic Acids Research,40(15):e115,2012.
[9] Bing Xia, Swapnil Bhatia, Ben Bubenheim, Maisam
Dadgar, Douglas Densmore, and J. Christopher
Anderson. Chapter five - developer’s and user’s guide
to clotho v2.0: A software platform for the creation of
synthetic biological systems. In Christopher Voigt,
editor, Synthetic Biology, Part B Computer Aided
Design and DNA Assembly, volume 498 of Methods in
Enzymology, pages 97 – 135. Academic Press, 2011.
17