Second-generation PLINK: rising to the challenge of larger and richer datasets

doi:10.1186/s13742-015-0047-8

PMC full text:

Gigascience. 2015; 4: 7.

Published online 2015 Feb 25. doi: 10.1186/s13742-015-0047-8

Copyright/LicenseRequest permission to reuse: Copyright © Chang et al.; licensee BioMed Central. 2015
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Table 4

–indep-pairwise runtimes (sec)

Parameters	Dataset	Machine	PLINK 1.07	PLINK 1.90	Ratio
50 5 0.5	synth1	Mac-2	701.3	0.63	1.1 k
		Mac-12	569.4	0.55	1.0 k
		Linux32-8	572.7	0.95	600
		Linux64-512	462	0.60	770
		Win32-2	1163.9	3.2	360
		Win64-2	1091.9	1.0	1.1 k
700 70 0.7	synth2	Mac-2	~120 k	31.9	3.8 k
		Mac-12	63.0 k	20.6	3.06 k
		Linux32-8	57.4 k	66.0	870
		Linux64-512	~120 k	26.4	4.5 k
		Win32-2	139.3 k	127.3	1.09 k
		Win64-2	~200 k	22.9	8.7 k
20000 2000 0.5	chr1	Mac-2	nomem	1520.1
		Mac-12	nomem	1121.7
		Linux32-8	nomem	4273.9
		Linux64-512	~950 k	1553.3	610
		Win32-2	nomem	4912.7
		Win64-2	nomem	1205.1
	1000g	Mac-2	nomem	20.5 k
		Mac-12	nomem	14.5 k
		Linux32-8	nomem	54.5 k
		Linux64-512	~13000 k	20.2 k	640
		Win32-2	nomem	64.5 k
		Win64-2	nomem	14.7 k

This command is used to select a set of genetic markers which are not too highly correlated with one another. The PLINK 1.9 implementation benefits from laziness (i.e. the correlation coefficient between a pair of variants is no longer computed when it is not needed by the main pruning algorithm) and bitwise operations.