created by Geraldine_VdAuwera
on 2014-02-12
Alright, the next release is going to be version 3.0. So what’s in it??
We’ll have a full overview ready for you in the next few days.
In the meantime, if you’re at AGBT-2014 working on your tan (lucky devil), one way to find out is to go see Mauricio Carneiro’s poster during the Thursday afternoon poster session, and ask him all about it (you’re welcome, MC — we know you like the attention).
If you’re not, here’s a copy of Mauricio’s poster, which features three of the top features in GATK 3.0. Because why should you miss out, on top of having to shovel snow all over again tomorrow (or whatever the applicable chore is in your neck of the woods) instead of drinking margaritas by the pool?
In case you’re wondering, the owl is the mascot of the nightly builds. He/she (we’re not sure; we respect its privacy) builds a fresh copy of the GATK every night with the day’s new developments that made it into the master branch.
agbt14, poster, conferences, gatk 3.0
Updated on 2014-02-12
From JiantaoShi on 2014-02-13
This is exciting! Looking forward to it.
From aeonsim on 2014-03-11
Hi
Were any of these performance improvements in the current version 3.0 of the Haplotype Caller? As I’ve been trying out the new single sample at a time approach (GVCF) following the docs and I’m seeing an estimated run time for a single 2.6Gb (Bovine) genome @ 24x coverage of ~70 hours on an AMD (Opteron 6380) 2.5GHz processor (-nct 6 cores for this test). Or does the GVCF mode prevent the optimisations from working?
The Genome Analysis Toolkit (GATK) v3.0-0-g6bad1c6, Compiled 2014/03/06 06:30:35 … INFO 15:40:22,637 HelpFormatter – Program Args: -R /home/aeonsim/refs/bosTau6.fasta -T HaplotypeCaller -I /home/projects/bos_taurus/damona/bams/H1B131204/NL288458773_H1B131204_L008_BQSR.bam -I /home/projects/bos_taurus/damona/bams/H2A131008-H1B131015/NL288458773_H2A131008_BQSR.bam -o chad.test.gatkUG.vcf.gz -D /home/aeonsim/refs/BosTau6_dbSNP138_NCBI.vcf.gz -nct 6 —emitRefConfidence GVCF —variant_index_type LINEAR —variant_index_parameter 128000
From ebanks on 2014-03-11
Sorry, no – the optimizations didn’t make it into the 3.0 release. We are targeting them for 3.1.