hammerlab / mhcflurry-icml-compbio-2016 Goto Github PK

Hi @timodonnell and @iskandr,

Clarification question:

A dataset of peptide-MHC affinities for n peptides and a alleles may be thought of as a n × a matrix
where peptide/allele pairs without measurements are missing values.

Using the dataset BD2009, it appears that you also filtered out peptides which occurred in at least three alleles and alleles with less than five measurements.

Did you measure percentage of NaN values you are imputing (i.e. how sparse is this matrix)? When doing the evaluation of imputation methods, what was the size of this (n,a) matrix?

Also, apologies for the simple question, but I'm slightly confused by the pre-print. Based on Kim et al, 2014, BD2009 contains 79 alleles, and 170 total datasets. Your BD2009 dataset has 106 alleles? And it was compiled into a single dataset?