I cannot detect a deletion which is seen by IGV

muratisbilen · 08-23-2017, 04:25 AM

Hi everybody,

I use samtools and bcftools 1.4.1 for variant calling. However, I cannot detect deletions for a specific position. In fact, when I visualize my data in IGV (by Broad Institute) I can see an obvious heterozygous deletion from T to no nucleotide (52% T, 45% - , 3% others) at 3294155th position of chr16 (hg19). However, samtools and bcftools cannot detect this deletion. What could be the reason for that? How can I modify my code shown below to call this deletion?

samtools mpileup -ABuvf ~/Desktop/Analysis/hg19/hg19.fa ${outpath}/sortedBowtieOut.sam -o ${outpath}/mpileup.vcf

bcftools call -v -c -O v -o ${outpath}/sortedBowtieOut_samtools_raw_variants.vcf ${outpath}/mpileup.vcf

Thanks, Kind Regards,

gringer · 08-23-2017, 04:30 AM

Samtools mpileup has its own INDEL correction, which may be influencing the results that you're seeing. You could try skipping this (option `-I´) and seeing if it changes the results.

Brian Bushnell · 08-23-2017, 10:43 AM

I'd suggest trying a different variant caller. The one I wrote (callvariants.sh in the BBMap package) works very well for calling indels from Illumina data, in terms of concordance with what you see in IGV.

muratisbilen · 09-05-2017, 04:43 AM

Thanks guys but neither of them worked. callvariants.sh gives an out of memory error :/

To test the ability of bcftools to detect INDELs, I generated a dummy data, which includes 100 forward and 100 reverse reads. Among these, 50% of the paired reads have a deletion on a specific position. bcftools still cannot detect such a pure deletion. How can I use/change parameters to make bcftools call such indels?

Regards,

finswimmer · 09-07-2017, 10:22 PM

Hello,

can you post a screenshot of IGV in this region? If I found it correctly, this is a position in a homopolymer run (11xT) which is typical not easy to sequence, align and discover a variant.

As Brian suggested try out another variant caller. My favorite is freebayes.

fin swimmer

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How to get deletion count from bam file in IGV?	nottsbio	Bioinformatics	6	07-22-2015 05:08 AM
What is deletion?	huanggm	General	6	09-25-2012 05:17 AM
IGV v2 performance much worse than IGV 1.5?	pmiguel	Bioinformatics	1	05-17-2012 08:05 PM
deletion mapping problem	StefanP	Bioinformatics	3	10-06-2011 04:20 AM
How to detect real deletion or gaps in sequencing projects	watashi	Bioinformatics	0	08-01-2008 02:15 PM

08-23-2017, 04:25 AM	#1
muratisbilen Junior Member Location: Kocaeli Join Date: Aug 2017 Posts: 2	I cannot detect a deletion which is seen by IGV Hi everybody, I use samtools and bcftools 1.4.1 for variant calling. However, I cannot detect deletions for a specific position. In fact, when I visualize my data in IGV (by Broad Institute) I can see an obvious heterozygous deletion from T to no nucleotide (52% T, 45% - , 3% others) at 3294155th position of chr16 (hg19). However, samtools and bcftools cannot detect this deletion. What could be the reason for that? How can I modify my code shown below to call this deletion? samtools mpileup -ABuvf ~/Desktop/Analysis/hg19/hg19.fa ${outpath}/sortedBowtieOut.sam -o ${outpath}/mpileup.vcf bcftools call -v -c -O v -o ${outpath}/sortedBowtieOut_samtools_raw_variants.vcf ${outpath}/mpileup.vcf Thanks, Kind Regards,

08-23-2017, 04:30 AM	#2
gringer David Eccles (gringer) Location: Wellington, New Zealand Join Date: May 2011 Posts: 838	Samtools mpileup has its own INDEL correction, which may be influencing the results that you're seeing. You could try skipping this (option `-I´) and seeing if it changes the results.

08-23-2017, 10:43 AM	#3
Brian Bushnell Super Moderator Location: Walnut Creek, CA Join Date: Jan 2014 Posts: 2,707	I'd suggest trying a different variant caller. The one I wrote (callvariants.sh in the BBMap package) works very well for calling indels from Illumina data, in terms of concordance with what you see in IGV.

09-05-2017, 04:43 AM	#4
muratisbilen Junior Member Location: Kocaeli Join Date: Aug 2017 Posts: 2	Thanks guys but neither of them worked. callvariants.sh gives an out of memory error :/ To test the ability of bcftools to detect INDELs, I generated a dummy data, which includes 100 forward and 100 reverse reads. Among these, 50% of the paired reads have a deletion on a specific position. bcftools still cannot detect such a pure deletion. How can I use/change parameters to make bcftools call such indels? Regards,

09-07-2017, 10:22 PM	#5
finswimmer Member Location: Europe Join Date: Oct 2016 Posts: 60	Hello, can you post a screenshot of IGV in this region? If I found it correctly, this is a position in a homopolymer run (11xT) which is typical not easy to sequence, align and discover a variant. As Brian suggested try out another variant caller. My favorite is freebayes. fin swimmer