rwdavies
commented
3 years ago Hi Lee, In general that is far fewer samples than I would recommend. I would start with K=2 and see how it performs, and then maybe for upwards of the 40 case, try K=2 or K=3. For filtration, without any other evidence which would suggest particular thresholds, I normally suggest filtering pretty much exclusively based on the INFO scores, e.g. keeping INFO>0.4 as the only filtration Best wishes, Robbie
HI Robbie, I had two questions： 1，For a relatively small sample (4-40 individuals), with ~1X per-sample，How to set parameters k and nGen(for 4 samples，k=4 and nGen=20)？ 2，How to set filter parameters after impution？ Thanks! Lee