Content uploaded by Craig Forest
Author content
All content in this area was uploaded by Craig Forest on Mar 28, 2017
Content may be subject to copyright.
Proceedings of the 2009 ASME International Mechanical Engineering Congress & Exposition
IMECE2009
November 13-19, 2009, Lake Buena Vista, Florida, USA
IMECE2009-13272
A LOW-COST, TWO-AXIS, PRECISION ROBOT FOR AUTOMATED
FLUORESCENCE IN-SITU HYBRIDIZATION ASSAYS
M.J. KULIK, D.S. SHENODA, C.R. FOREST
Department of Mechanical Engineering, Georgia Institute of Technology
Atlanta, GA, USA
ABSTRACT
Genetics research often relies on experiments that
require repetitive, time-consuming handling of small
volumes of liquid (1 mL) and biomass (10-20 μL) such as
fluorescence in-situ hybridization (FISH), β-galactosidase
staining, immunohisto chemistry, skeletal and tunel
assays. Often manual, these experiments are time
intensive and error-prone. We report on the design,
fabrication, and testing of a low-cost, two-axis, precision
robot for FISH assays on whole mice embryos. The robot
can complete 20 successive embryo immersions in
unique isothermal solutions in minutes for 6 samples.
Repeatability of the orthogonal axes is 66 and 214 μm,
near the measurement uncertainty limit and sufficient for
operation. Accuracy is achieved by systematic error
compensation. Low-cost and precision are obtained
using design and manufacturing techniques and
processes, resulting in a cost of 15% of comparable
instruments (e.g., InsituStain, Intavis Bioanalytical
Instruments). This design demonstrates a simple,
automated platform to perform a typically manual
experimental genetics technique.
INTRODUCTION
Genetics research is replete with experiments that
require repetitive, time-consuming handling of small
volumes of liquid (1 mL) and biomass (10-20 μL). Manual
labor in handling these volumes leads to increased errors
in accuracy and repeatability, increased costs and time.
For example, fluorescence in-situ hybridization (FISH) is
a three day experimental procedure that primarily
involves washing samples such as embryos in different
solution baths for different periods of time, ranging from
5-120 min. Each of these three days involves an eight
hour protocol to physically transport samples within 12
mm diameter baskets between solution baths.
Several automation solutions have been developed
specifically for FISH including the InsituStain [1] and the
Automated Immunohistochemical and In Situ
Hybridization Assay Formulations Instrument [2]. In both
cases, these instruments are prohibitively expensive and
complex for many laboratories and experiments,
approaching $60,000. Further, the samples are held
fixed while solutions are washed over them, which can
damage tiny samples such as mouse embryos (14 μL) of
interest in this work. More generally, a variety of flexible,
multi-axis robots have been developed to automate
chemical and biological laboratory practices (e.g., [3-5]).
In this work, we focus on low-cost and precision for a set
of laboratory tasks that involve delicate samples being
repeatedly immersed in solutions without damage, an
unmet laboratory automation need. There is a need to
deliver automated, precision controlled (100 µm), low-
cost (<$8,000), low-volume (µL-mL) instrumentation.
DESIGN
Mechanical
We report on the design and fabrication of a low-cost,
two-axis, precision robot for automated fluorescence in-
situ hybridization assays, shown in Fig. 1. We designed
the instrument to be capable of translating and immersing
up to six baskets containing embryos between a total of
20 rows of wells amongst five well plates (2 mL/well, 24
wells/plate, Corning Inc.) at room temperature.
As shown in Fig, 1, both axes are driven by stepper
motors with 800 micro-counts/revolution (Danaher
Motion, CT Series, CTP10ELF10MMA00). The horizontal
(x-axis) utilizes a lead screw transmission (M6, 1 mm
pitch) resulting in 1.25 μm resolution, 500 mm range.
The vertical assembly (see Fig. 1 , Fig. 2) utilizes a rack
and pinion drive with 100 μm resolution, 20 mm range.
2 Copyright © 2009 by ASME
To reduce cost, we have attempted to minimize part
count, keep the form factor compact and rectilinear, use
uniform fasteners (M3), use off-the-shelf parts (e.g., lead
screw all-thread), use the same motors, bearings (nylon
bushings), and ground steel guide rods for both axes, and
mass align the well trays with a pair of orthogonal, spring-
loaded constraint plates rather than individual positioners.
Further, the structure of the instrument is entirely
comprised of extruded aluminum sections (80/20 Inc., 25-
2525) with associated fasteners and brackets. Low cost,
versatile manufacturing processes such as waterjet
cutting, milling, and turning were used throughout to
reduce fabrication costs.
Figure 1. Photograph of a low-cost, two-axis, precision
robot for automated fluorescence in-situ hybridization
assays.
Figure 2. Photograph of the vertical assembly within the
robot instrument for raising and lowering baskets into
wells for hybridization assays.
Baskets (not shown) are held in the basket tray which
is constrained by spring-loaded ball-nose plungers within
the vertical assembly. These 14 mm diameter baskets
are porous with a 100 µm mesh screen. Embryos are
loaded into the basket manually.
The wells within the well plates are filled with solutions
using pipettes. Thus the volume accuracy and precision
are controlled by the pipetting technique, typically 1%.
The volumes required are less than comparable
instruments in which solutions are washed over the
samples.
Software/Electronics
Automated control is performed using LabView to
easily control the stepper motors through a programming
interface and hardware data I/O boards. The simple
graphical user interface (GUI), shown in Fig. 3, requires
the user to specify the amount of time that the baskets
dwell at each row of solution baths. The axes’
acceleration, deceleration and velocity are also
programmable.
The motors are controlled by individual drivers (Copley
Controls, STP-075-07) via a PCI card (Copley Controls,
CAN-PCI-02) installed in a computer. The control
algorithm was calibrated to the spacing between adjacent
wells and adjacent well plates.
Figure 3. LabView interface, or GUI, for robot control.
in
p
ut: dwell time/well
outputs: dwell time
remainin
g
and well number
3 Copyright © 2009 by ASME
RESULTS AND DISCUSSION
We fabricated the instrument as shown in Fig. 1 and
performed a series of experiments to characterize its
performance. The x-axis and z-axis were programmed to
travel at 5 mm/sec and 10 mm/sec respectively across
the range of motion. The instrument moves between
successive wells as intended, requiring a few minutes for
a full run with 1 sec dwell time/well. The positioning
accuracy required is determined in the x-axis by the
difference in the radii of the well and basket, 2 mm, and in
the z-axis by the depth of the solution and height of the
embryo. For full immersion, this z-axis accuracy
requirement is 1 mm. These accuracies were achieved
by calibration of the well and basket positions prior to the
start of an experiment. The repeatability of positioning in
the axes was determined by repeatedly (20 trials) moving
to a well position from a “home” position and measuring
the actual basket location. Uncertainties in this
repeatability measurement were determined by
repeatedly measuring basket position without moving it
between measurements. The stated repeatability and
uncertainty is the standard deviation of these
measurements. The uncertainty is attributable to errors in
positioning of a hand-held digital caliper with 10 µm
resolution. Table 1 shows the measured repeatabilities
and uncertainties for both axes.
Table 1. Repeatability and uncertainty of measurement for
the instrument axes
x-axis
(horizontal)
(µm)
z-axis
(vertical)
(µm)
repeatability 66 214
uncertainty 65 89
Measurement of the repeatability in the x-axis of the
instrument is limited by the uncertainty—it is, at worst, 66
µm, which is sufficient for the operations required. In the
z-axis, measurement of the repeatability is not limited by
the uncertainty. The repeatability in this axis is 214 µm,
largely due to the low-cost rack and pinion transmission.
The total instrument cost is 15% of comparable
instruments (e.g., InsituPro VSi, Intavis Bioanalytical
Instruments). Thus this design demonstrates a simple,
automated platform to perform a typically manual, time
intensive, error-prone, experimental genetics technique.
Future work will include independent well plate
temperature control for more experimental freedom as
well as a web-accessible interface to further reduce cost
and footprint by eliminating the need for a display. The
current design meets the accuracy and repeatability
requirements, while automating the process in a low-
volume, low-cost design.
ACKNOWLEDGMENTS
We are grateful to Ashley Michelle Moroz, Raj Shah,
and John Near for their help and work in the preliminary
design of the robot for the Capstone Design course in the
GW Woodruff School of Mechanical Engineering at the
Georgia Institute of Technology.
The contributions and mentoring of Dr. Tamara
Caspary from the Department of Human Genetics at
Emory University School of Medicine regarding FISH
assays as well as the genesis of the project are greatly
appreciated.
The assistance and generous support of the Precision
Biosystems Laboratory at the Georgia Institute of
Technology are also greatly appreciated.
REFERENCES
[1] INTAVIS Bioanalytical Instruments, Nattermannallee 1,
50829 Koeln Germany.
[2] Town, P. et al. “Automated immunohistochemical and
in situ hybridization assay formulations.” U.S. Patent
6,855,552. 15 Feb 2005.
[3] Saitoh, S., and Yoshimori, T., 2008, “Fully Automated
Laboratory Robotic System for Automating Sample
Preparation and Analysis to Reduce Cost and Time in
Drug Development Process,” Journal of the Association
for Laboratory Automation, 13(5), pp. 265-274.
[4] Najmabadi, P., and Goldenberg, A. A., 2005, “A
scalable robotic-based laboratory automation system for
medium-sized biotechnology laboratories,” Proceedings
of the 2005 IEEE Conference on Automation Science and
Engineering, IEEE-CASE 2005, pp.166-171.
[5] Ward, K. B., Perozzo, M. A., and Zuk, W. M., 1987,
“Automatic preparation of protein crystals using
laboratory robotics and automated visual inspection,”
Journal of Crystal Growth, 90(1-3), pp. 325-339.
