Haplotypes-phenotype association analysis using PLINK

[Aroosha Raja]

Hi,

I have already assembled haplotypes from the set of SNPs for above 160,000 samples. Now I have a text file in which I have haplotypes in columns and Samples in row and the values are 0 and 1. 0 represents the absence of haplotype in sample whereas 1 represents presence of haplotype in sample. I have above 300 clinical features for the samples under study. I want to associate each haplotype with each clinical feature using Logistic regression model while correcting for Sex and Age. How can I do it in Plink? 

[Christopher Chang]

In general, to perform analysis with PLINK, you need to convert your data into one of the formats PLINK is capable of importing.  With PLINK 2.0, see https://www.cog-genomics.org/plink/2.0/input#vcf and the sections below it on the webpage.

In your case, it should be easiest to convert to .haps format (https://www.cog-genomics.org/plink/2.0/formats#haps ), which can be imported with --haps; just duplicate each of your haplotype columns.  (PLINK is mostly geared toward analysis of diploid genotypes, so we're pretending that each sample has two identical haplotypes here.)