Minimal input to validate plumbing of vector short read pipelines about pipelines

[tnguyensanger]

@hardingnj @alimanfoo

I am considering sharing the lanelet crams for samples: AV0148-C and AV0079-C. Their indel realigned bams both have average coverage of ~2.5x, and thus both failed sampled QC. Both have very few called genotypes. We can definitely use these samples to test plumbing of IRODS import --> alignment --> UnifiedGenotyper.

Do you suspect the lack of genotype calls will cause these samples to crash the new phasing pipeline?

For testing just plumbing of alignment -->UnifiedGenotyper, we can downsample those crams to just a single chromosome.

[hardingnj]

Thanks Thuy.

(below applies to shapeit2 only. I am unsure of the behaviour of shapeit4)

Shapeit2 reports an error when all of the genotypes in a dimension are missing. ie a sample with all missing genotypes, or a genotype missing in all samples.

Phasing just 2 samples both with high missingness, would almost certainly meet this condition. That being said, it's very possible that a GT call may be missing in all samples in the whole set, so may well need consideration regardless and not just in this minimal case.

We hit this mainly when we were running single samples against a reference panel. In that case, any missing genotypes were problematic.

[tnguyensanger]

I've decided to leave the samples AV0148-C and AV0079-C as is without downsampling, since they are 3 orders of magnitude smaller than the existing samples we've provided: AN0131-C and AB0252-C. Per-lanelet bams of ~1.7GB vs 2.5MB.

[alimanfoo]

Thanks both. FWIW I don't think we need to worry about the phasing pipeline at this time, the approach to validation there will need to be entirely different anyway.

…

On Fri, 29 May 2020 at 23:04, tnguyensanger ***@***.***> wrote: I've decided to leave the samples AV0148-C and AV0079-C as is without downsampling, since they are 3 orders of magnitude smaller than the existing samples we've provided: AN0131-C and AB0252-C. Per-lanelet bams of ~1.7GB vs 2.5MB. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#25 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFLYQXVBDNT3YNUXWS7B2DRUAWQHANCNFSM4NLGD5CA> .

-- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health Big Data Institute Li Ka Shing Centre for Health Information and Discovery University of Oxford Old Road Campus Headington Oxford OX3 7LF United Kingdom Phone: [phone] or [phone] Email: [email protected] Web: http://a <http://purl.org/net/aliman>limanfoo.github.io/ Twitter: @alimanfoo <https://twitter.com/alimanfoo> Please feel free to resend your email and/or contact me by other means if you need an urgent reply.

[tnguyensanger]

Now that we've started doing batch processing of multiple samples at once, we need more plumbing samples. We have about 2 really small samples. We need at least 4.

@magnusmanske We can use AV0148-C and AV0079-C for quick batch processing testing. That allows to have 1 sample per batch. I'll try to find 2 samples so that we can have at least 2 samples per batch.