now finds flexbounds/relpos using a "representative" clonal sequence (instead of aggregating over CF)
now makes separate directories for the "partition" FASTA's
removed the +/- 2 bounds fuzz (for short D gene matches, this leads to errors!)
The "representative" stuff was inspired by Laura's messy QA255.067-Vh data, which had tons of uncertainty and D gene matches of length 4 with tons of D genes aligned all over the place, and this caused the flexbounds/relpos to be inconsistent with each other.
The "representative" stuff was inspired by Laura's messy
QA255.067-Vhdata, which had tons of uncertainty and D gene matches of length 4 with tons of D genes aligned all over the place, and this caused the flexbounds/relpos to be inconsistent with each other.