Convert Impute2 phased output files to mach format

Question

[kindlychung]

I am trying to convert some files from impute2 phased output (haps files),

Most of the lines look like this:

        --- SNP BP A1 A2 ID1HAP1   ID2HAP2 ...
        --- rs62224609 16051249 T C 0 0 0 0 0 0 0 1 0 0 1
        --- rs62224610 16051347 G C 0 1 0 1 1 0 0 1 0 0 1
        --- rs143503259 16051453 A C 0 0 0 0 0 0 0 1 0 0
        --- rs192339082 16051477 C A 0 0 0 0 0 0 0 0 0 0
        --- rs79725552 16051480 T C 0 0 0 0 0 0 0 0 0 0 0
        --- rs141578542 16051497 A G 0 1 0 1 1 0 0 1 0 0
        --- rs201906224 16051722 TA T 0 0 0 0 0 0 0 0 0 0
        --- rs2843213 16051882 C T 0 0 0 0 0 0 0 0 0 0 0
        --- rs4965031 16052080 G A 0 0 0 0 0 0 0 0 0 0 0
        --- rs6518413 16052239 A G 0 0 0 0 0 1 0 0 0 0 0

Question 1: Does 0 stand for allele 1 and 1 for allele 2? For example, in this line:

            --- rs6518413 16052239 A G 0 0 0 0 0 1 0 0 0 0 0

Does 0 stand for A and 1 for G here?

Question 2: Some lines contain multiple characters in the A1 or A2 column, what do those characters stand for? (I am guessing indel, but not sure.)

    --- SNP BP A1 A2 ID1HAP1 ID2HAP2 ...
    --- chr22:16078656 16078656 G GTGTC 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
    --- rs199998412 16134558 ATAACT A 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    --- rs201164934 16151190 TGCCTA T 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    --- rs199662619 16166919 ATATTTTCTGCACATATT A 0 0 0 0 0 0 0 0 0 0 0 0
    --- rs200691780 16197677 TAAAG T 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    --- chr22:16231367 16231367 G GAGAA 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
    --- rs201020033 16368171 CAGAG C 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    --- chr22:16380919 16380919 A AAAAT 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
    --- rs141841004 16432988 GTACT G 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    --- rs200126408 16459572 TATATATAG T 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

For example, in this line:

    --- chr22:16078656 16078656 G GTGTC 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

what does GTGTC stand for?

 Here is a typical phased output from mach:

        232->232 HAPLO1 TTGACCCCGATGTGTTAAGACCGTATCACTCCACTCTTCAAGTCGGGATTGTC
        232->232 HAPLO2 TGACCCCCAACGTGCCAAGACCGTGCTGACCAGTCCTTTACACCGAAATTATT
        2921->2921 HAPLO1 TTGAACCCAACGTGCCAATGCTATGCTACCCAGCTTCCCAAGCCGAAGTTG
        2921->2921 HAPLO2 TGGCCCCCAATGTGCCAATGCTATGCTACCCAGTCCTCCGAGCTGAAGTTG
        3370->3370 HAPLO1 TTGAACCCGACGTGTTAAGACCGTATCACTCCATCCTTCAAGTCGGGACTA

For each individual, there are two lines, one line for each haplotype. The third column contains genotype data, one letter for each SNP.

So, question 3: in case I want to convert impute2 phased output to mach phased output, how should I represent the mutations more complicated than a bi-allelic polymorphism?

Answer 1

[Maxime Lamontagne]

Question 1: I think you are right. To be sure, how many columns do you have? If it's equal to the number of samples, each number is a genotype. If you have twice the number of samples, each number is an allele.

Question 2: Insertion / Deletion. Take one SNP and search it on UCSC. Ex: rs199998412 is a deletion (-/TAACT).

Question 3: 1) Why transform IMPUTE2 output to MaCH output? 2) MaCH output are number between 0 and 2. A1A1 in IMPUTE2 will be 0 in MaCH, A1A2 in IMPUTE2 will be 1 in MaCH and A2A2 in IMPUTE2 will be 2 in MaCH.

[kindlychung]

Thanks! About 3(2), MaCH also has a phased output format, in which you can use 1,2,3 and 4 to represent A, C, G and T. But I don't know how to handle indels here.