PDA

View Full Version : GEDmatch one-to-one comparison... what do these terms actually mean?



therrien.joel
09-27-2017, 02:20 AM
I could take an educated guess, but I'd like to know if someone understands these terms and what they technically mean. i.e. what impact do they have on the match results?

Size (in SNPs) of Mismatch Evaluation window

and

Size (in SNPs) of Mismatch-Bunching limit

therrien.joel
09-30-2017, 02:19 AM
Maybe to get the ball rolling I can describe how I would interpret these parameters...

Size of mismatch evaluation window: I'd guess this just means how "wide" of a scope should the evaluation be done over. Since this is typically defaulted to the same as the minimum SNP threshold, then it suggests that the smallest acceptable segment would be evaluated all at once. Something larger probably uses this as a sliding window so that it is less data intensive. A quick test where I make this value larger than the minimum threshold did eliminate a few small segments where making it smaller than the default added more small segments. This is at least consistent with this interpretation of the window concept.

Size of mismatch-bunching limit: I'd guess based on playing around with this one is that it limits how close mismatching SNPs can get.

So, it seems that if one is willing to play around in the weeds of small segments, you can be more conservative by increasing the limit on the mismatch-bunching limit. The evaluation window does not seem to do much except that increasing it and not entering a number for the bunching limit will increase that parameter by default too.