Hey everyone,
I'm attempting to do a large fimo run over a big set of TF motifs (a combination of 3 motif databases) on specifically footprint locations.
I'm using the --no-qvalue, --text, and the --skip-match-sequence options and piping the output to a file. Depending on how much ram I give fimo, it is able to get further and further through the TF database, but it's RAM usage constantly increases until it is killed by the kernal. In my mind, with those options there should be nothing retained between TF searches in that each instance that passes threshold is printed. I've even set my max-stored-scores to a tiny number and it doesn't seem to help.
Is there an option that I'm missing that can further reduce RAM usage? Should I just cut my motif database into smaller chunks and run each separately? Is this a known bug with FIMO?