[general question]gatk, dindel, IGV

[sirmark]

Hi all!
I'm newbie to analyzing NGS data and I'm not a biologist.

I'm very confused, and I have some question and hope that you can give me simple answer:

1th
Dindel is the GATK algorithm for indel calling ? If not, what do GATK use for indel calling and what is GATK algorithm for indel calling ?

2th:
IGV is a genomic viewer, that can show how shortreads alignment to reference genome, e.g.:


IGV use samtools to do counting, and other analysis on NGS data ?
if not: are samtools and IGV two tool indipendent ?

Sorry for my silly question...

[bruce01]

1: Dindel is a standalone software to call indels: http://www.sanger.ac.uk/resources/software/dindel/. GATK uses UnifiedGenotyper to call indels (as well as SNPs).

2:IGV does not use Samtools afaik. You need to specify 'other analysis'. If you are here you should know at least what analysis you need to do.

[vdauwera]

Yeah, you're confused alright, but we're here to help ;-)

1) Dindel and GATK are separate things -- As far as I know Dindel is a single tool focused on calling indels, whereas GATK is a toolbox with many tools including some that do indel calling.

You can find details on GATK tools here: http://www.broadinstitute.org/gatk/guide/

In a nutshell, GATK has two separate tools to call indels: UnifiedGenotyper and HaplotypeCaller. HaplotypeCaller is newer and more sensitive for indels than UnifiedGenotyper. Eventually it will replace UG completely.

2) IGV is mainly a genome viewer. IGV is developed at the same place as GATK (Broad Institute) so they work well together. I work in the group that develops GATK, and we use IGV all the time to look at the sequence data we work with.

Samtools is a different toolbox, developed elsewhere. It includes many useful tools, some of which we use to complement the GATK. It also includes a variant caller called mpileup. The alogrithms used by mpileup and by GATK's UnifiedGenotyper are similar because Heng Li, the author of mpileup, helped with the original design of the UGs algorithm, if I remember correctly.

Since you are new to NGS analysis, I recommend you watch the introduction videos we provide here:

http://www.broadinstitute.org/gatk/guide/events?id=2038

I hope that will help you understand better.

If you decide to work with GATK, feel free to ask questions in our support forum here:

http://gatkforums.broadinstitute.org/

Good luck!

[sirmark]

1: Dindel is a standalone software to call indels: http://www.sanger.ac.uk/resources/software/dindel/. GATK uses UnifiedGenotyper to call indels (as well as SNPs).

2:IGV does not use Samtools afaik. You need to specify 'other analysis'. If you are here you should know at least what analysis you need to do.

for other analysis I mean always indel calling.
Do IGV do indel calling ?


About GATK why is it possible to use DINDEL with -glm option ?

thanks for help!

[vdauwera]

No, IGV does not do indel calling.

It is no longer possible to use -glm DINDEL with newer versions of GATK. Much older versions did have that mode because the original indel calling model was inspired by the standalone tool DINDEL. But now the model is completely different.

You should really use the latest version of GATK in order to get quality results for indel calling.

[sirmark]

Yeah, you're confused alright, but we're here to help ;-)

1) Dindel and GATK are separate things -- As far as I know Dindel is a single tool focused on calling indels, whereas GATK is a toolbox with many tools including some that do indel calling.

You can find details on GATK tools here: http://www.broadinstitute.org/gatk/guide/

In a nutshell, GATK has two separate tools to call indels: UnifiedGenotyper and HaplotypeCaller. HaplotypeCaller is newer and more sensitive for indels than UnifiedGenotyper. Eventually it will replace UG completely.

2) IGV is mainly a genome viewer. IGV is developed at the same place as GATK (Broad Institute) so they work well together. I work in the group that develops GATK, and we use IGV all the time to look at the sequence data we work with.

Samtools is a different toolbox, developed elsewhere. It includes many useful tools, some of which we use to complement the GATK. It also includes a variant caller called mpileup. The alogrithms used by mpileup and by GATK's UnifiedGenotyper are similar because Heng Li, the author of mpileup, helped with the original design of the UGs algorithm, if I remember correctly.

Since you are new to NGS analysis, I recommend you watch the introduction videos we provide here:

http://www.broadinstitute.org/gatk/guide/events?id=2038

I hope that will help you understand better.

If you decide to work with GATK, feel free to ask questions in our support forum here:

http://gatkforums.broadinstitute.org/

Good luck!

thank you very very much, I'm confused less!

I actually use IGV to see alignment of my BAM file to reference genome hg19.
To indel calling, I visualize manually each entire chromosome.

I'm finding a way in IGV to find the indel position for entire my bam file and for all chromosome not manually. Is it possible in IGV or is it necessary to use tool as: gatk, samtools ?

Last question: what's the role of igvtools ?

[sirmark]

thank you very very much, I'm confused less!

I actually use IGV to see alignment of my BAM file to reference genome hg19.
To indel calling, I visualize manually each entire chromosome.

I'm finding a way in IGV to find the indel position for entire my bam file and for all chromosome not manually. Is it possible in IGV or is it necessary to use tool as: gatk, samtools ?

Last question: what's the role of igvtools ?

[vdauwera]

I strongly suggest you use a dedicated indel calling tool rather than try to identify indels manually. As far as I know GATK's HaplotypeCaller gives the best results for indels, but others may have different opinions (since I work on GATK, I may be biased).

IGVtools are utilities that can convert some formats or count some basic statistics on your dataset. They are useful for these tasks but are not a replacement for variant calling tools.

[sirmark]

ok! I understand.

in your opinion, in case of pooled samples, which tools are the best ?
I'm finding an article with compared to evaluate this.

[vdauwera]

HaplotypeCaller cannot handle pooled samples yet, so for now you should use UnifiedGenotyper.

[sirmark]

just some questions :
dindel is only for indel or also for snp ?

is there a reference/publication that compare different snp calling algorithms ?