SNP/Indel analysis and homopolymer errors

gavin.oliver · 11-08-2011, 02:45 AM

Hi,

I am wondering what the general consensus is on SNP/Indel detection from resequencing data generated by homopolymer error-prone platforms like 454 and Ion torrent?

Off the top of my head I theorize that SNP detection should generally not be a problem but analysis of Indels would be. Am I correct in this assumption?

Are there any general anlysis approaches that attempt to compensate for these problems?

Thanks in advance.

nickloman · 11-08-2011, 03:30 AM

Hi Gavin, generally speaking that's true assuming sufficient depth of coverage. Your aligner needs to be produce high quality gapped alignments when dealing with Ion Torrent or 454 data, which is why people tend to use aligners like TMAP, Newbler gsMapper, Novoalign, SSAHA2 or BWA-SW for these datasets.

This is a good treatment from EdgeBio of this specific problem comparing Ion Torrent and Illumina variant calling: http://www.edgebio.com/blog/?p=442

Generally speaking you can filter your SNPs by increasing thresholds on metrics such as variant frequency and proximity to homopolymers to gain a "trusted" set of variants. This means you are trading sensitivity for specificity.

gavin.oliver · 11-09-2011, 12:52 AM

Thanks a lot for the input Nick - it's a great help.

gavin.oliver · 11-10-2011, 07:57 AM

Nick, can I just follow this up?

Between MiSeq and Ion Torrent - which do you see having the greater clinical application (generally)?

Also, more specifically are the homopolymers likely to be an issue in such applications?

Cheers in advance.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Using solexa to correct 454 homopolymer errors	coldturkey	Bioinformatics	21	01-26-2016 02:39 AM
Pileup - Filter SNP/Indel linked to 454 homopolymer	bardou	Bioinformatics	0	06-30-2010 08:17 AM
correcting homopolymer run errors	454andSolid	Bioinformatics	4	05-02-2010 09:40 AM
Calculation of Homopolymer errors	pcg	454 Pyrosequencing	0	04-12-2010 09:15 AM
454 homopolymer errors or????	ian Adams	454 Pyrosequencing	9	12-02-2008 02:46 AM

11-08-2011, 02:45 AM	#1
gavin.oliver Senior Member Location: uk Join Date: Jan 2010 Posts: 110	SNP/Indel analysis and homopolymer errors Hi, I am wondering what the general consensus is on SNP/Indel detection from resequencing data generated by homopolymer error-prone platforms like 454 and Ion torrent? Off the top of my head I theorize that SNP detection should generally not be a problem but analysis of Indels would be. Am I correct in this assumption? Are there any general anlysis approaches that attempt to compensate for these problems? Thanks in advance.

11-08-2011, 03:30 AM	#2
nickloman Senior Member Location: Birmingham, UK Join Date: Jul 2009 Posts: 356	Hi Gavin, generally speaking that's true assuming sufficient depth of coverage. Your aligner needs to be produce high quality gapped alignments when dealing with Ion Torrent or 454 data, which is why people tend to use aligners like TMAP, Newbler gsMapper, Novoalign, SSAHA2 or BWA-SW for these datasets. This is a good treatment from EdgeBio of this specific problem comparing Ion Torrent and Illumina variant calling: http://www.edgebio.com/blog/?p=442 Generally speaking you can filter your SNPs by increasing thresholds on metrics such as variant frequency and proximity to homopolymers to gain a "trusted" set of variants. This means you are trading sensitivity for specificity.

11-09-2011, 12:52 AM	#3
gavin.oliver Senior Member Location: uk Join Date: Jan 2010 Posts: 110	Thanks a lot for the input Nick - it's a great help.

11-10-2011, 07:57 AM	#4
gavin.oliver Senior Member Location: uk Join Date: Jan 2010 Posts: 110	Nick, can I just follow this up? Between MiSeq and Ion Torrent - which do you see having the greater clinical application (generally)? Also, more specifically are the homopolymers likely to be an issue in such applications? Cheers in advance.