help with polyphen output

[kjaja]

I have a list of variants and want to know if the variants are predicted to be deleterious or benign. When I did that using Polyphen2 and uploaded my variants as batch query, I got multiple possibilities for the prediction. For example, for the variant chr11:47306106 G/A, below is the Polyphen output, which one to pick? And what does this mean

o_snp_id o_acc o_pos o_aa1 o_aa2 snp_id nt1 nt2 prediction
chr11:47306106.GA.uc001ner.1 Q8WXG6 716 G E chr11:47306106.GA.uc001ner.1 G A possibly damaging
chr11:47306106.GA.uc001nes.1 Q8WXG6-4 716 G E chr11:47306106.GA.uc001nes.1 G A benign
chr11:47306106.GA.uc001net.1 Q8WXG6-2 716 G E chr11:47306106.GA.uc001net.1 G A benign
chr11:47306106.GA.uc001neu.1 Q8WXG6-5 716 G E chr11:47306106.GA.uc001neu.1 G A possibly damaging
chr11:47306106.GA.uc001nev.1 Q8WXG6-6 716 G E chr11:47306106.GA.uc001nev.1 G A benign
chr11:47306106.GA.uc001new.2 Q8WXG6-3 716 G E chr11:47306106.GA.uc001new.2 G A possibly damaging
chr11:47306106.GA.uc001nex.2 Q8WXG6-7 716 G E chr11:47306106.GA.uc001nex.2 G A possibly damaging
chr11:47306106.GA.uc001nez.2 Q8WXG6-5 716 G E chr11:47306106.GA.uc001nez.2 G A possibly damaging
chr11:47306106.GA.uc009yln.1 B5MEE5 716 G E chr11:47306106.GA.uc009yln.1 G A benign

acc pos aa1 aa2 nt1 nt2 prediction
Q8WXG6 716 G E G A possibly damaging
Q8WXG6-4 716 G E G A benign
Q8WXG6-2 716 G E G A benign
Q8WXG6-5 716 G E G A possibly damaging
Q8WXG6-6 716 G E G A benign
Q8WXG6-3 716 G E G A possibly damaging
Q8WXG6-7 716 G E G A possibly damaging
Q8WXG6-5 716 G E G A possibly damaging
B5MEE5 716 G E G A benign

[iadzhubey]

The protein encoded by this gene (MADD) has 8 different isoforms and the position of your variant happens to overlap with all of them. Hence you got predictions for all sequences, which may differ, and do differ slightly in your case. You may decide to go with the prediction outcome for "canonical" isoform (Q8WXG6), or select more suitable isoform based on your knowledge of specifics of the biology of the process you study, or even try to "average" all predictions for all isoforms. There were 2 "benign" and 3 "possible damaging" predictions, the latter being not a strong effect indicator, hence overall, your variant is more probable to be benign.

The last isoform (B5MEE5) comes from an unvalidated EMBL transcript and can be ignored.

For details, see:

http://www.uniprot.org/uniprot/Q8WXG6

http://www.uniprot.org/uniprot/B5MEE5