Hail (@hailgenetics) - Twitter Profili | Zamantika Mersobahis Locabet

Sabitlenmiş Tweet

Hail@hailgenetics·7 Kas

Hey all! We are running a survey to solicit broad-based feedback from users & *non-users* of Hail Query. We'd like to hear from anyone who analyzes sequencing or genotype-chip datasets! All questions optional, fully anonymous, & takes <=10min! forms.gle/ucgu9h35UEkB68…

English

0

2

3

3K

Hail retweetledi

Konrad Karczewski@konradjk·27 Kas

CHARR operates only on homozygous alternate sites and scales very well (“cost per 1M samples” might be my new favorite metric):

English

1

3

1K

Hail retweetledi

Kaitlin Samocha@ksamocha·13 Kas

Now that the dust has settled on the gnomAD v4 release, which hopefully many of you have already checked out, I wanted to take this week to thank many of the members of the team who made this possible. First up this week is the amazing production team. 1/7

English

2

13

85

10K

Hail@hailgenetics·14 Kas

@ngsstudent @ksamocha I suspect 10-30TiB for 1M samples, given certain assumptions, but we haven't actually tested it at 1M samples. Even so, I think SVCR-VCF is mostly useful for *interchange*. IMHO (@danking00 ), the days of analyzing sequencing data with text tools is gone. Just too much data 3/3

English

0

35

Hail@hailgenetics·14 Kas

@ngsstudent @ksamocha PVCF is dead. We're finishing up a preprint that describes the "Scalable Variant Call Representation" and a proposed VCF implementation called "SVCR-VCF". SVCR-VCF represents both samples and alleles sparsely, allowing for a much smaller representation 2/3

English

1

2

149

Hail retweetledi

Kaitlin Samocha@ksamocha·4 Kas

Julia Goodrich is now up to talk about QC in "many, many exomes and genomes". Earlier this week, we released gnomAD v4 -- 76k genomes and 731k exomes. Close to 1 BILLION variants. But how do we ensure data quality at this scale? #ASHG2023 #ASHG23

English

1

4

20

3.1K

Hail retweetledi

Hail@hailgenetics·7 Kas

Hey all! We are running a survey to solicit broad-based feedback from users & *non-users* of Hail Query. We'd like to hear from anyone who analyzes sequencing or genotype-chip datasets! All questions optional, fully anonymous, & takes <=10min! forms.gle/ucgu9h35UEkB68…

English

0

2

3

3K

Hail retweetledi

Genome Aggregation Database@gnomad_project·8 Kas

Please take a few minuets to provide feedback to the #hail team!

Hail@hailgenetics

Hey all! We are running a survey to solicit broad-based feedback from users & *non-users* of Hail Query. We'd like to hear from anyone who analyzes sequencing or genotype-chip datasets! All questions optional, fully anonymous, & takes <=10min! forms.gle/ucgu9h35UEkB68…

English

0

2

1.8K

Hail@hailgenetics·4 Kas

@ngsstudent @ksamocha Yeah this was a possibly poor naming choice on our part.

English

1

0

59

aaaa@ngsstudent·4 Kas

@hailgenetics @ksamocha Ok! As I was confused because it was with old version of hail.

English

1

0

1

59

Hail@hailgenetics·4 Kas

@ngsstudent @ksamocha The VDS is a new sparse representation we developed. It stores both samples and alleles sparsely (ref blocks for sample sparsity, “local” allele indices for alleles). Under the covers, it’s a pair of matrix tables: one containing reference data and one containing variant data.

English

1

2

655

aaaa@ngsstudent·4 Kas

@ksamocha VDS or matrix ? 🤔

Deutsch

1

0

65

Hail retweetledi

Wolfgang M. Pernice@joungMax·2 Kas

This is big! #gnomAD has enabled countless breakthroughs in human genetics to date. Still, especially for rare genetic diseases, we face countless diagnostic riddles. Can't wait to see the impact this update will have (and of the tech behind it @hailgenetics). Thanks for this 🙏

Genome Aggregation Database@gnomad_project

The #gnomAD team is proud to announce the release of gnomAD v4! The v4 dataset includes 730,947 exomes & 76,215 genomes, which is ~5x larger than the v2 & v3 releases combined, & includes nearly 120K indivs of non-European genetic ancestry broad.io/gnomad_v4 #ASHG23 (1/11)

English

0

1

5

607

Hail@hailgenetics·1 Kas

[1] Assuming reasonably reblocked GVCFs; bigger GVCFs of course drive cost up!

English

0

108

Hail@hailgenetics·1 Kas

With the current version of Hail the cost to produce a VDS from GVCFs, is down to 0.005 USD per exome [1]. The VDS is also compact, costing about 0.0005 USD per month per exome to store in Google Cloud Storage (~25MiB per exome). 2/3

English

2

0

2

221

Hail@hailgenetics·1 Kas

The released set has ~800k samples, but we started with a 955,000 sample call set! *Staggeringly* large! We developed the Scalable Variant Call Representation (SVCR) and implemented it as the Hail VDS to enable this nearly 1M sample callset. 1/3

Genome Aggregation Database@gnomad_project

The #gnomAD team is proud to announce the release of gnomAD v4! The v4 dataset includes 730,947 exomes & 76,215 genomes, which is ~5x larger than the v2 & v3 releases combined, & includes nearly 120K indivs of non-European genetic ancestry broad.io/gnomad_v4 #ASHG23 (1/11)

English

1

6

32

4.7K

Hail@hailgenetics·1 Kas

And, of course, all that functionality is publicly available as MIT licensed software! Check out the docs at: hail.is/docs/0.2/vds/i… Huge congrats to the gnomAD team for another massive release! 3/3

English

0

2

140

Hail retweetledi

Duncan Palmer@astheeggeggs·31 Mar

We’re excited to announce that our analysis of genetic dominance in the UK Biobank genotype data is out in @ScienceMagazine (science.org/doi/10.1126/sc…) (1/12)

English

1

47

146

44.3K

Hail@hailgenetics·16 Mar

The @broadinstitute is accepting applications for its Artist in Residence program! Come hang out with the scientists and make some art! broadinstitute.org/artist-residen…

English

0

183

Hail retweetledi

Alicia Martin@genetisaur·24 Oca

Way to go @ZanKoenig and @itsnotmeron for developing scalable cloud-based tutorials for common genomics analyses: github.com/atgu/hgdp_tgp/…. We have also released phased haplotypes for phasing/imputation: gs://gcp-public-data--gnomad/resources/hgdp_1kg/phased_haplotypes

English

1

5

31

2.7K

Hail

Keşfet