we discussed this as a potential way to speed up VCF writing in simgenotype
it seems like a promising idea because the number of flags that we output scales linearly with the number of samples, so it's probably responsible for a good portion of the file IO when the number of samples is large
we discussed this as a potential way to speed up VCF writing in simgenotype
it seems like a promising idea because the number of flags that we output scales linearly with the number of samples, so it's probably responsible for a good portion of the file IO when the number of samples is large
we think we can test this in parallel with #99