2
$\begingroup$

I would like to compare GATK with graph-based variant callers. I have seen the Fast and accurate genomic analyses using genome graphs paper from SevenBridges and the paragraph tool by Illumina, though it does not look like folks use their platform or it is not open-sourced (in the case of SevenBridges's GRAF).

The idea is to see if graph-based approaches detect better both SNPs and structural variants compared with existing methods. Additionally, I believe that if there exists some novel graph-based aligner/assembler one could use the metadata downstream to identify variants.

$\endgroup$
2
  • $\begingroup$ Not sure about variant callers, but since you asked about aligners, you should check out vg. $\endgroup$
    – user438383
    Feb 9 at 9:48
  • $\begingroup$ @user438383 thanks, it's worth mentioning WhatsHap, GraphAligner, and Minigraph as well. $\endgroup$
    – 0x90
    Feb 9 at 11:48
4
$\begingroup$

Yes, there are. There are some suggestions in the comments (VG, WhatsHap, GraphAligner, Minigraph).

However, to be clear, the current default variant calling algorithm used in the GATK (the haplotype caller) is a graph-based algorithm. It constructs a De Bruijn graph from mapped reads for local assembly around variants. Other tools that also follow this strategy include Platypus and freebayes.

You might also consider multi-color genome assemblers. For example, Cortex and Corticall use colored De Bruijn graphs to aggregate reads from multiple genomes and call variants.

$\endgroup$
2
  • $\begingroup$ I think there is a difference between using de Bruin graph and represent a genome as a graph. The colored graph assemblers are interesting. Thank you. $\endgroup$
    – 0x90
    Feb 11 at 7:22
  • $\begingroup$ I completely agree. $\endgroup$
    – winni2k
    Feb 11 at 7:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.