Precomputed Genome level hasfrag

[ejarmand]

Hi @muntakimrafi wanted to create a an issue relevant to our discussion on Blue sky

Basically it would be really useful to provide files with precomputed pairwise leakage for all genomic elements. I imagine every ~200 bp or so, for popular references (hg38, mm10, T2T, GCRm39).

Ideally there could also be a tiled whole genome pipeline.

If it's useful I'm happy to contribute some development time to parts of this task.

[muntakimrafi]

Hello @ejarmand, continuing from our discussion on Blue sky

This would be an enourmous amount of calculation. But I think we should do this if we were do a create the least possible leakage free splits for the genome.

The first step would be to run blastn genome wide. I am thinking of creating mutliple databases (1 per chromosome) and running blastn_array modules (for each chromosome as query set) (@bkiyota just pushed it to the repo). This way we can modularize the computation and divide it among multiple interested parties.