Biology Reference
In-Depth Information
the plate, and “reverse” sequencing dye is added to the bottom
half of the plate.
4
The fourth and fi fth steps are “ethanol precipitation” and “elution.”
respectively. Both of these chemical processes are designed to remove
excess dye and other contaminants. All fi ve of these steps are far from
automated: each involves careful laboratory work, testing, and judgment
calls. In picking, the agar plates are “visually checked” for approximate
accuracy of counts and spacing; the 384-well plates are checked for a
“hazy” appearance (indicating a growing colony) before sequencing; so-
lutions used in the TempliPhi steps are checked each morning (randomly
selected plates are weighed to ensure that correct amounts of reagent
are being added).
5
This is not a straightforward sequence of steps, but
rather an interlocking set of checks and tests that require knowledge of
the process chemistry.
It is in the sixth step, “detection,” that the samples disappear wholly
into the virtual. And it is only in this step that the samples are introduced
into the sequencing machines. About 120 Applied Biosystems 3730 de-
tectors, each costing hundreds of thousands of dollars, sit together in
a large single room—each is about 6 feet high and has a footprint of
roughly 3 feet by 2 feet. Because of the cost of the machines and the
speed at which they process samples, detection is both the rate-limiting
and the cost-limiting step. As such, the detectors are operated twenty-
four hours per day, seven days per week. Inside the machines, the dyes
attached in step three are excited by a laser and detected by a CCD
camera. In typical detector output, the four possible nucleotide bases
are represented by lines of four different colors on a chart (red, green,
blue, black).
These colored graphs are known as the “raw sequence traces.” They
are stored by the sequencing machines, but they are not yet “fi nished”
sequence. As the raw sequence traces are generated by the detectors, two
computational steps are automatically performed. The fi rst is known as
“base calling” and is usually done using software called “Phred.” Phred
analyzes the raw sequence trace images and, based on the height and
position of each peak, makes a “call” as to what the correct base is for
each position along the sequence. In other words, it converts an image
fi le (the four-colored chart) into a text string consisting of the familiar
As, Gs, Ts, and Cs. It also assigns a quality score to each base, refl ect-
ing its confi dence in the call that it makes. The second automated step
is “sequence assembly.” Recall that in the very fi rst stages of the se-
quencing process, before the DNA was spliced into plasmids, the entire
