Analysis commons, a team approach to discovery in a big-data environment for genetic epidemiology

Brody, Jennifer A.; Morrison, Alanna C.; Bis, Joshua C.; O’Connell, Jeffrey R.; Brown, M. R. W.; Huffman, Jennifer E.; Ames, Darren; Carroll, Andrew; Conomos, Matthew P.; Gabriel, Stacey; Gibbs, Richard A.; Gogarten, Stephanie M.; Gupta, Namrata; Jaquish, Cashell E.; Johnson, Andrew D.; Lewis, Joshua P.; Liu, Xiaoming; Manning, Alisa K.; Papanicolaou, George; Pitsillides, Achilleas N.; Rice, Ken; Salerno, William; Sitlani, Colleen M.; Smith, Nicholas L.; Heckbert, Susan R.; Laurie, Cathy C.; Mitchell, Braxton D.; Vasan, Ramachandran S.; Rich, Stephen S.; Rotter, Jerome I.; Wilson, James G.; Boerwinkle, Eric; Psaty, Bruce M.; Cupples, L. Adrienne

doi:10.1038/ng.3968

Cited by 95 publications

(77 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 1 shows how reference panel size has increased over the years due to projects such as the International HapMap Project 9 , the 1000 Genomes Project (1000G) 10 , the UK10K Project 11 , and the Haplotype Reference Consortium (HRC) 12 . Soon larger reference panels will become available from the Trans-Omics for Precision Medicine (TOPMed) program 13 and the 100,000 Genomes Project 14 , both of which will exceed 100,000 high-coverage whole genome sequenced samples. Further ahead, as sequencing data on all 500,000 participants of the UK Biobank 2 becomes available, this will be used as an even larger reference panel.…”

Section: Introductionmentioning

confidence: 99%

Genotype imputation using the Positional Burrows Wheeler Transform

Rubinacci

Delaneau

Marchini³

2019

Preprint

View full text Add to dashboard Cite

Genotype imputation is the process of predicting unobserved genotypes in a sample of individuals using a reference panel of haplotypes. Increasing reference panel size poses ever increasing computational challenges for imputation methods. Here we present IMPUTE5, a genotype imputation method that can scale to reference panels with millions of samples. It achieves fast and memory-efficient imputation by selecting haplotypes using the Positional Burrows Wheeler Transform (PBWT), which are used as conditioning states within the IMPUTE model. IMPUTE5 is 20x faster than MINIMAC4 and 3x faster than BEAGLE5, when using the HRC reference panel, and uses less memory than both these methods. IMPUTE5 scales sub-linearly with reference panel size. Keeping the number of imputed markers constant, a 100 fold increase in reference panel size requires less than twice the computation time.

show abstract

Section: Introductionmentioning

confidence: 99%

Genotype imputation using the Positional Burrows Wheeler Transform

Rubinacci

Delaneau

Marchini³

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…The value of larger reference panels has led to steadily increasing reference panel size, with the largest reference panels to date having thousands or tens of thousands of samples. [6][7][8][9][10][11][12] Ongoing, large-scale projects, such as the Trans-Omics for Precision Medicine (TopMed) program, 13 and the National Human Genome Research Institute's Centers for Common Disease Genomics, 14; 15 are expected to produce reference panels with more than one hundred thousand individuals.…”

Section: Introductionmentioning

confidence: 99%

A one penny imputed genome from next generation reference panels

Browning

Zhou

Browning

2018

Preprint

129

161

View full text Add to dashboard Cite

Genotype imputation is commonly performed in genome-wide association studies because it greatly increases the number of markers that can be tested for association with a trait. In general, one should perform genotype imputation using the largest reference panel that is available because the number of accurately imputed variants increases with reference panel size. However, one impediment to using larger reference panels is the increased computational cost of imputation. We present a new genotype imputation method, Beagle 5.0, which greatly reduces the computational cost of imputation from large reference panels. We compare Beagle 5.0 with Beagle 4.1, Impute4, Minimac3, and Minimac4 using 1000 Genomes Project data, Haplotype Reference Consortium data, and simulated data for 10k, 100k, 1M, and 10M reference samples. All methods produce nearly identical accuracy, but Beagle 5.0 has the lowest computation time and the best scaling of computation time with increasing reference panel size. For 10k, 100k, 1M, and 10M reference samples and 1000 phased target samples, Beagle 5.0's computation time is 3x (10k), 12x (100k), 43x (1M), and 533x (10M) faster than the fastest alternative method. Cost data from the Amazon Elastic Compute Cloud show that Beagle 5.0 can perform genomewide imputation from 10M reference samples into 1000 phased target samples at a cost of less than one US cent per sample.Beagle 5.0 is freely available from https://faculty.washington.edu/browning/beagle/beagle.html.

show abstract

“…In recent years, large-scale whole-exome sequencing and whole-genome sequencing (WGS) data have been generated, such as those in the Exome Sequencing Project 1 , the UK10K project 2 and the ongoing NIH NHLBI Trans-Omics for Precision Medicine (TOPMed) WGS program 3 , providing unprecedented opportunities to investigate low-frequency variants (minor allele frequency [MAF] between 1% and 5%) and rare variants (RVs; MAF < 1%) in association with complex diseases and traits. However, WGS-based association analysis of complex traits remains a tremendous challenge due to the large number of RVs, many of which are non-traitassociated neutral variants.…”

Section: Introductionmentioning

confidence: 99%

FunSPU: a versatile and adaptive multiple functional annotation-based association test of whole-genome sequencing data

Wei

2018

Preprint

View full text Add to dashboard Cite

Despite ongoing large-scale population-based whole-genome sequencing (WGS) projects such as the NIH NHLBI TOPMed program and the NHGRI Genome Sequencing Program, WGS-based association analysis of complex traits remains a tremendous challenge due to the large number of rare variants, many of which are non-trait-associated neutral variants. External biological knowledge, such as functional annotations based on ENCODE, may be helpful in distinguishing causal rare variants from neutral ones; however, each functional annotation can only provide certain aspects of the biological functions. Our knowledge for selecting informative annotations a priori is limited, and incorporating non-informative annotations will introduce noise and lose power. We propose FunSPU, a versatile and adaptive test that incorporates multiple biological annotations and is adaptive at both the annotation and variant levels and thus maintains high power even in the presence of noninformative annotations. In addition to extensive simulations, we illustrate our proposed test using the TWINSUK cohort (n=1,752) of UK10K WGS data based on six functional annotations: CADD, RegulomeDB, FunSeq, Funseq2, GERP++, and GenoSkyline. We identified genome-wide significant genetic loci on chromosome 19 near gene TOMM40 and APOC4-APOC2 associated with low-density lipoprotein (LDL), which are replicated in the UK10K ALSPAC cohort (n=1,497). These replicated LDL-associated loci were missed by existing rare variant association tests that either ignore external biological information or rely on a single source of biological knowledge. We have implemented the proposed test in an R package "FunSPU".

show abstract

Analysis commons, a team approach to discovery in a big-data environment for genetic epidemiology

Cited by 95 publications

References 11 publications

Genotype imputation using the Positional Burrows Wheeler Transform

Genotype imputation using the Positional Burrows Wheeler Transform

A one penny imputed genome from next generation reference panels

FunSPU: a versatile and adaptive multiple functional annotation-based association test of whole-genome sequencing data

Contact Info

Product

Resources

About