Peaktrace vs ABI KB basecaller

Use this forum for general bioinformatics questions, or questions regarding genomics, proteomics, etc.

Moderators: mdfenko, Abhijeet Bakre, strom

Peaktrace vs ABI KB basecaller

Postby SeqInfant » Aug 07 2014 12:59 pm

Dear members,

What do you think about a DNA sequencing service provider generate "outstanding" chromatogram using software ie Peaktrace ( http://www.nucleics.com/peaktrace/ )??? Usually ABI KB basecaller will show chromatogram peak resolution decrease after 700-800bp eventhough it can give 1000+bp base call for one sequence read/However using Peaktrace it is possible to see a chromatogram that still show very sharp and nice and "high resolution" even near 1000bp...... Somebody think a DNA sequencing facility had done a better job than the other because it is using Peaktrace to "produce" the chromatogram while others are using ABI original software to make base call. What do you think??? Hope to hear your opinions about this... :D :D :D
SeqInfant
newcomer
newcomer
 
Posts: 2
Joined: Aug 02 2014 12:47 pm

Re: Peaktrace vs ABI KB basecaller

Postby folli » Dec 05 2014 4:05 am

I worked for a sequencing provider which pays a lot of money to use peaktrace.

Since I don't believe that the costumer wants to go through the trouble of running their own peaktrace analysis, I think it's only in the interest of the client to get the most out of the data.
folli
newcomer
newcomer
 
Posts: 2
Joined: Dec 05 2014 3:45 am

Re: Peaktrace vs ABI KB basecaller

Postby r.rosati » Dec 05 2014 10:39 am

It can be possible to extract the original data from the .abi file, if Peaktrace doesn't erase it. Abi files often carry, dereferenced (so practically as dead weight), quite a few partial data. I've used the abifpy Python script recently, and I was surprised at both the amount of info that you can read from an ABI file, and how quite a significant amount of the file size is actually dereferenced, and only useful if you want to play the "ABI archeology game".
r.rosati
ModSquad
ModSquad
 
Posts: 2133
Joined: Nov 04 2002 10:23 am
Location: Brazil

Re: Peaktrace vs ABI KB basecaller

Postby DanielTillett » Aug 25 2015 5:56 am

As the person who actually wrote PeakTrace I can tell you that the improvement that PeakTrace offers is real. While not magic or perfect, it is basically just a better basecaller in exactly the same way that KB is a better basecaller than the old ABI basecaller. Of course I am biased, but any sequencing provider should be using the best tools available and if you are doing Sanger sequencing there is no better basecaller than PeakTrace. Rather than complain if it is “fair” why not try it (it is free to try) and see for yourself [1].

For those interested in the nitty-gritty details the validity of PeakTrace has been checked and is described in the PeakTrace Whitepaper [2]. Interestingly, as far as I know ABI has not published the validation of KB.


1. PeakTrace Online - this is the free version.
2. PeakTrace Whitepaper
Last edited by DanielTillett on Aug 27 2015 4:56 pm, edited 1 time in total.
DanielTillett
technician-in-training
technician-in-training
 
Posts: 7
Joined: Aug 25 2015 3:04 am
Location: Australia

Re: Peaktrace vs ABI KB basecaller

Postby DanielTillett » Aug 25 2015 5:52 pm

r.rosati wrote:It can be possible to extract the original data from the .abi file, if Peaktrace doesn't erase it. Abi files often carry, dereferenced (so practically as dead weight), quite a few partial data. I've used the abifpy Python script recently, and I was surprised at both the amount of info that you can read from an ABI file, and how quite a significant amount of the file size is actually dereferenced, and only useful if you want to play the "ABI archeology game".


Yes the ABI file contains the raw data. PeakTrace extracts this and reprocesses it before doing a new basecall. The reason PeakTrace offers better basecalling is because it does a better job of processing the raw data than KB. Other pure basecallers like phred and TraceTuner are limited to just re-bacecalling the processed peak data. This is why they can’t do any better than KB while PeakTrace can.

Yes the ABI file format is an interesting file format. It contains features that go back to ABI 373. The same information is repeated multiple times across the file (from memory the record for data duplication is 6 times). The really frustrating part is that some software will use the information from one part of the file, while another time it will use the same information contained at another location. This means that you really need to keep most (all) of the file if you want to maintain compatibility with downstream applications. ABI’s software is one of the worst offenders for doing this with different versions of the same software (e.g. Sequence Analysis) using different values in what appears to be an almost random fashion.
DanielTillett
technician-in-training
technician-in-training
 
Posts: 7
Joined: Aug 25 2015 3:04 am
Location: Australia


Return to General Bioinformatics

Who is online

Users browsing this forum: No registered users and 1 guest