phased and unphased genotype results differ ?

Question

[CrazyB]

I was plowing through two sets of VCF outputs - one with genotype phased and the other unphased. To my surprise, a few genotype data seem different. Any idea??

CHR POS RS REF ALT SAMPLE1 SAMPLE2 SAMPLE3
PHASED
chr11 118956652 . C A 0|0 1|0 0|1
chr11 118956655 . A AG 1|0 0|0 1|0
chr11 118956656 . A ACACCCCGGCTGCAGCTGCCGAGGGCCTCACTCACCTTC 0|0 1|0 0|1

UNPHASED
chr11 118956652 . C A 0/0 0/0 0/1
chr11 118956655 . A AG 0/0 0/0 0/1
chr11 118956656 . A ACACCCCGGCTGCAGCTGCCGAGGGCCTCACTCACCTTC 0/0 0/0 0/1

[Pierre Lindenbaum]

you didn't describe the tool and the options used ....

[Pierre Lindenbaum]

Please validate or comment your previous questions:

• a question about some "basic numbers" for rna-seq experiments
• understanding Cuffdiff comparison (or a general understanding of pair-wise v muti-groups comparision)
• Regarding deleted SNPs (in dbSNP) due to mapping or clustering errors found in VCF file
• how to find SNP positions (for non-bioinformaticians)
• any tools for annotating "canonical splice site variants"
• ...

[CrazyB]

sorry - not sure what I should do about my prior questions ... e.g. "validating" my questions ?? Please advise (and will do it). Thanks.

[Pierre Lindenbaum]

When you ask a question, you'll see a grayed-out checkbox with a "Accept!" mouseover on all answers given to your question. When you accept an answer, the poster get reputation points, but more important, people with problems similar to yours will easily find a solution which at least worked before.

[CrazyB]

@ Pierre, to clarify, should I check "accept" out of courtesy for all responses or only when I believe the provided answer helps with the question?

Answer 1

[Zev.Kronenberg]

Phasing software can change genotypes. This usually happens for variant with low genotype likelihoods.