LDkit: a parallel computing toolkit for linkage disequilibrium analysis

You Tang,Aoxue Wang,Helong Yu,Zhuo Li,Yao Zhou,Chao Wang,Yuxin Liu

doi:10.1186/s12859-020-03754-5

Abstract

BackgroundLinkage disequilibrium (LD) analysis is broadly utilized in genetics to understand the evolutionary and demographic history and helps geneticists identify genes associated with interested inherited traits, such as diseases. There are some tools for linkage disequilibrium analysis either in a local or online way; however, there has been no such tool supporting both graphical user interface (GUI) and parallel computing.ResultsWe developed a GUI software called LDkit for LD analysis, which supports parallel computing. The LDkit supports both variant call format (VCF) and PLINK ‘ped + map’ format. At the same time, users could also just analyze a subset of individuals from the whole population. The LDkit reads the data by block and then paralleled the computation process by monitoring the usage of processes. Assessment on the Human 1000 genome data showed that when paralleled with 32 threads, the running time was reduced to less than 6 minutes from ~77 minutes using the chromosome 22 dataset with 1,103,547 SNPs and 2504 individuals.ConclusionsThe software LDkit can be effectively used to calculate and plot LD decay, LD block, and linkage disequilibrium analysis between a site and a given region. Most importantly, both graphical user interface (GUI) and stand-alone packages are available for users’ convenience. LDkit was written in JAVA language under cross-platform support.

Highlights

Linkage disequilibrium (LD) analysis is broadly utilized in genetics to understand the evolutionary and demographic history and helps geneticists identify genes associated with interested inherited traits, such as diseases
Most of the population genetic analyses, such as detection of sites under selection, estimation of divergence time between different populations, positional cloning, and genome wide association studies (GWAS) are based on linkage disequilibrium (LD) theory, which has been investigated over 100 Years [1]
LD decay analysis To assess the performance of LDkit on LD decay analysis, we tested the software using the chromosome 22 dataset from the Human 1000 genome

Summary

Results

We developed a GUI software called LDkit for LD analysis, which supports parallel computing. The LDkit supports both variant call format (VCF) and PLINK ‘ped + map’ format. Users could just analyze a subset of individuals from the whole population. The LDkit reads the data by block and paralleled the computation process by monitoring the usage of processes. Assessment on the Human 1000 genome data showed that when paralleled with 32 threads, the running time was reduced to less than 6 minutes from ~77 minutes using the chromosome 22 dataset with 1,103,547 SNPs and 2504 individuals

Conclusions

Background

Results and discussion