search

szhan / onekg_analysis

Evaluation of genotype imputation methods using the unified genealogy dataset
MIT License
0 stars 0 forks source link

Run `tskit.lshmm` using BEAGLE like HMM parameters #24

Closed szhan closed 1 year ago

szhan commented 1 year ago

See #21. The formula to determine mismatch and switch probabilities takes number of haplotypes as input. I'm doing two runs, the first time by taking the number of reference and target haplotypes (which is what BEAGLE does) and the second time by taking the number of reference haplotypes.

szhan commented 1 year ago

The number of wrongly imputed alleles is still quite high, as shown below comparing imputed alleles and true alleles.

Screen Shot 2023-07-03 at 7 54 09 AM

szhan commented 1 year ago

Also, below is an example where a switch shouldn't happen, but it did. It is the first switch in the traceback path going from right to left.

Screen Shot 2023-07-03 at 7 50 29 AM