Closed chrisquince closed 4 years ago
Although with the current method of detecting overlap this will require us to extract COG subgraphs for all bins so maybe better to do something where we thread the relevant COG ORFs instead?
OK this has now been implemented so closing.
When we check COG subgraphs for overlaps we only compare to other MAGs not all bins. This means that if the contaminating COG is not in a MAG we miss it. We should filter the MAG COGs by overlaps with all bins.