paperKB
coga / coga-kb
Help
Sign in

Chunk #53 — Findings — Performance comparisons

Source
Second-generation PLINK: rising to the challenge of larger and richer datasets.
Embedded
yes

Text

Ghz Intel Celeron G1610T processor (2 cores) and 8GB RAM running 64-bit Windows 8.“synth1” refers to a 1000 sample, 100000 variant synthetic dataset generated with HAPGEN2 [37], while “synth1p” refers to the same dataset after one round of –indep-pairwise 50 5 0.5 pruning (with 76124 markers remaining). For case/control tests, PLINK 1.9’s –tail-pheno 0 command was used to downcode the quantitative phenotype to case/control.“synth2” refers to a 4000 case, 6000 control synthetic dataset with 88025 markers on chromosomes 19-22 generated by resampling HapMap and 1000 Genomes data with simuRare [38] and then removing monomorphic loci. “synth2p” refers to the same dataset after one round of –indep-pairwise 700 70 0.7 pruning (with 71307 markers remaining).“1000g” refers to the entire 1092 sample, 39637448 variant 1000 Genomes project phase 1 dataset [39]. “chr1” refers to chromosome 1 from this dataset, with 3001739 variants. “chr1snp” refers to chromosome 1 after removal of all non-SNPs and one round of –indep-pairwise 20000 2000 0.5 pruning (798703 markers remaining). Pedigree information was not added to these datasets before our tests.