Automatic Sequencing by Capillary Electrophoresis

Tom Vink tvink at xs4all.nl
Thu Aug 19 05:39:16 EST 1999

On 18 Aug 1999 13:46:49 -0700, John Barlow <john_q_barlow at hotmail.com>

>Here is some specific information on the result of a recent test on the Beckman 
>CEQ 2000.  Another user (from Europe) reported exactly the same problem to me.  
>I hope that Beckman Project Managers read this group and that they will try to 
>get their programmers to work harder.  Present software is simply inadequate 
>and possibly was never shown to power users before releasing to the public.
>We definitely need a good instrument to compete with ABI products.  So far, 
>after the demise of the ALF and the fallout of Amersham with Visible Genetics 
>over the crappy quality of their sequencers, the CEQ 2000 may be the last hope. 
> You all know by now that the LiCor instruments have very vocal, albeit rare 
>supporters and the final word on LiCor is difficult to come by.
>By the way, the higher signal that we get with the ABI 377 than with the CEQ 
>2000 can be attributed to the BigDye technology.  It is probably the best thing 
>since sliced bread.
>We did some more testing on the CEQ.  Specifically, we sequenced about forty
>435 bases long PCR products.  Next we did the same thing that we do with the
>results from ABI 377 XL - we transferred the trace files to a PC running the
>latest version of SeqMan II from the LaserGene (DNAStar) suite.
>In SeqMan, we loaded the traces (in SCF format for the CEQ and in the ABI
>format from the 377), trimmed ends using the Medium option, and assembled
>the contig.  Next, we scanned the contig for ambiguous positions.
>The results from the Beckman were, unfortunately, quite disastrous.  First, we
>get a much higher (10 fold difference) signal with the ABI 377.  Next, the
>basecalling on the Beckman is very bad, just like we have observed it last
>March and May.  The problem is that when a position is ambiguous, the ABI
>basecalling will assign an "N".  Rarely, the ABI basecaller would fail and
>not call a base at all.  In our hands, this happens only on one strand and can 
>be corrected by comparison with the complementary strand in SeqMan.  On
>the other hand, the CEQ basecaller will frequently assign a wrong base where
>the ABI will show an "N".  In addition, the CEQ basecaller missed about 5%
>of the positions.  Proof reading with the CEQ data alone would not be
>possible, because of the low signal.  We had about 4 people testing these
>traces and all decided against the CEQ in favor of the 377.
>In article <7pa4qf$ggf at net.bio.net>, srlasky at cellworks.mbt.washington.edu 
>>"Stephen R. Lasky" wrote:
>>> I would like to refer you to the post made by John barlow 07/21/1999
>>> where he gives a good review he recieve in response to which sequencer
>>> to buy.
>>> srl
>>Boy, that didn't make any sense..  Try:  See the post by John Barlow
>>7/21/99.  He reviews the responses he received on the Beckman.
>>Stephen R. Lasky, Ph.D.                 #
>>University of Washington                #
>>Department of Molecular Biotechnology   #
>>srlasky at u.washington.edu                #

We recently also had a decline of the signal and attributed that to
the quiality of the formamide. The low signal could be elevated by
raising the inject time from 1 to 5 min. The bad base calling is
probably a result of the low signal you obtained . When we have a high
signal, the sequence calling is near perfect. Check your formamide or
deionize it again. Ions present in the formamide will compete with the
DNA during the loading process,  so also your DNA should be very clean
and desalted.

Greetings Tom

More information about the Autoseq mailing list

Send comments to us at biosci-help [At] net.bio.net