Grch38 And Reference Alignment

Question

[aniketd86]

I came across a slide share about the new GRCh38 assembly and its ALT_LOCI assemblies. (http://www.slideshare.net/vaschn/agbt2014schneider)

Question: 1. Are there any "ALT Loci aware" aligners out there ?

1. If so, then how are accommodating these ALT LOCI while aligning reads to the reference genome ? How do they chose between primary assembly and Alternate Loci ?

Some references: This Biostar thread (Applying patches to GRCh assembly) talks about patches/Alts during alignment as well as about the aligner - srprism (ftp://ftp.ncbi.nlm.nih.gov/pub/agarwala/srprism)

Nice overview of ALT LOCI in GRCh38 (Will GRC38/HG20 be a multiple sequence reference genome?)

thanks in Advance,

Aniket

[Mary]

Good question. There was a recent interview that had some nice details about the ALT features and other aspects of the new assembly: http://www.bio-itworld.com/2013/4/22/church-on-reference-genomes-past-present-future.html

Answer 1

[deanna.church]

There are two 'ALT loci aware' (or multi-allelic reference) aware aligners that I know about.

• srprism (ftp://ftp.ncbi.nlm.nih.gov/pub/agarwala/srprism)
• ibwa (https://github.com/genome/ibwa)

Neither are published (AFAIK) but srprism is in preparation. In general, I think they both need to be tested more. It is also unclear how other tools along the tool chain (variant calling) will respond- and they will likely also need modification to take advantage of the full assembly.

[aniketd86]

Deanna, thanks for the prompt answer ! Do you know of any variant callers making use of the multi-allelic reference.( future versions of GATK maybe) ? Also, how do you think variants on primary and alt sequences would be represented in a VCF file [Flag in the INFO col or list the same variant twice ] ?

[Sam]

I have also used the GSNAP aligner, seems to be able to account for the alt loci too and it has been published: http://bioinformatics.oxfordjournals.org/content/26/7/873.abstract

However, it does seems like it is focusing more on RNA alignment.

You can find the programme here: http://research-pub.gene.com/gmap/

[deanna.church]

I know of no variant callers that have been modified to handle this. sorry. :(