DNAseq: POP5 on 3700

Phillip San Miguel pmiguel at purdue.edu
Thu Mar 30 11:08:06 EST 2000


Stephanie Johnson wrote:

> >[...]When switching to v1.1 and POP5, were does base calling start (in
> >bases away from the primer)? On the runs I have seen, the first about
> >70 or so bases are lost due to mobility shifts. Our single 3700 is
> >now running with v1.1, but we still use POP6. Has the problem been
> >solved or is it anyhow only roumors?
> >
> >Thanks, Bernd
> [...]Have there been any further conclusions about what this mobility
> shift
> is?  I have been trying to optimize our new (and only) 3700, and
> regularly see these mobility shifts in the first 70 bases.  When I
> called tech support, they stated that they had never heard of this
> problem and were not sure of what could cause it.
> [...]

Stephanie,
    I have no idea how to solve this problem--but it really exists. (At
least for water-loaded samples. Maybe it doesn't happen with formamide?) It
seems only to be a problem with products terminating in dye-term "A"
shifting forward with respect to the other dye-terminated products. In an
example I am looking at now[1], the 13th base (an "A") after the end of the
priming site runs right on top of the 14th base. By base 50 the A's are
shifted only 1/2 or 1/4th of a "base space" forward. By eye, there is no
problem calling bases past about 50 from the end of the primer but the
spacing confuses the ABI basecaller somewhat. Oddly, even 200 bases out,
clumps of 3 "A's" in a row look to me a little shifted forward. But the
basecaller has no problem here.
    One possibility is that the algorithm used to correct for the differing
dye mobilities is calibrated wrong [2]. This doesn't seem likely because
later in the run the A-terminated strands are shifted correctly. But the dye
mobility correction may be a linear correction. That is, if a short (ie,
close to the primer) G-terminated strand is 1 unit faster (than the other
terminator bases)  then a long G-terminated strand should also be 1 unit
faster. But maybe this is not the case for A-terminated products in POP5.
Then the data extracting/basecalling software may not be sophisticated
enough to correct for the A-terminated strands behavior in POP5.
    Generally all the problems are in the vector sequence, so it doesn't
really effect me much. So I haven't been complaining about it. But if you
need close to the priming site, it would be problematic.
Phillip San Miguel
Purdue University Genomics Core Facility


[1]I should add that I'm looking at the mobility-corrected chromatogram
window using ABI's Sequencing Analysis 3.6 program. If I look at the "raw"
data and expand it by hitting "control-+" and then scrolling until I find
the start of the peaks, it looks like the "A's" *are* being shifted back
some early in the read in the mobility-corrected window. Just not enough.
[2] I did do a new spectral when I shifted to POP5 (v. 1.1).

---






More information about the Autoseq mailing list