Has anyone used this DNA sequencing product before? - (Mar/06/2006 )

Hi Daniel,
We get our samples sequenced at the Australian Genome Research Facility and get our results back via FTP (we get an .ab1 file and a sequence file) but I don't know what kind of ABI machine they have.
If I understand this correctly, can I simply just upload my .ab1 files and get a better base-call? If it's just a matter of an optimised base-call algorithm, wouldn't a place like the AGRF already implement it? Maybe the traces we get from them are as good as it gets? How would I know?

---

Hi Daniel,

do you have any experience with bisulfite modified direct cycle sequencing and your product? I uploaded tzo sequences and for one I got no results (capillary weak?) for the other I got a good result, as KB basecalling seemed to work in your software's hands - I never had luck with ABI's KB directlw, always used different basecallers... Is it a problem for you software that you have C/T peaks?

Kr¨mel

---

The AGRF has two ABI 3730xl DNA sequencers (well they did last time I was there). So yes you can just upload your traces and get a better base call.

LongTrace is more than just a better basecaller. What is does is reprocess the raw signal in the trace (this is why it only works with .ab1 files) and then gives this to KB for the base call. This is why LongTrace generate a standard KB base called trace. My recommendation is try it on a few traces and see how it goes - if your trace is improved by LongTrace then it is very obvious. I should mention that LongTrace makes good traces better not failed trace good. The reason for this is reprocessing requires real signal and if your traces has failed then their is no signal only noise.

As for why a facility like AGRF has not implemented LongTrace it is sadly because they don't really care about the quality of the traces they give their customers. They get paid the same if the trace is good or bad, in fact they will earn more if the trace is bad because you will need to run more reactions. Thankfully some facilities care about quality

Cheers

Daniel

---

Hi Krumel

LongTrace doesn't work with traces that have two peaks at the same location. It is an assumption of the algorithms that there is only one peak at each location so it won't work with trace that have C/T peaks. Having said this the software can handle a small number of double peaks so if the trace doesn't have too many then it will be able to be improved by LongTrace. Was the trace that was improved a bisulfite trace?

The result "capillary weak" means that the signal at the end of the trace has reached noise level. LongTrace can't improve such traces as there is no peak signal to improve (only noise). We have tried to avoid the problem that KB often shows where noise is turned into fake peaks

Cheers

Daniel

---

I see. I tried it (see attachment), the electropherogramm looks great, but the necessary information (i.e. the C signal indicating >50% methylation in two spots (between 110 and 120 in the lower original lane, corresponding to 140-150 in the upper long trace lane) is lost after the long trace procedure.
But for other sequencing questions I will surely give it a try! And maybe nucleics comes up with a long trace meth - tool in the near future

Krümel

---