Low-complexity filtering currently discards molecules for which the entropy of the full cloneID is below 1.0. There are some decisions that could be made in other ways that could potentially lead to better filtering results.
Use a different threshold
Change what is done with deleted/missing bases (0 and -). As suggested by @acorbat, we could compute entropy on the cloneID without them and rescale entropy.
Low-complexity filtering currently discards molecules for which the entropy of the full cloneID is below 1.0. There are some decisions that could be made in other ways that could potentially lead to better filtering results.
0and-). As suggested by @acorbat, we could compute entropy on the cloneID without them and rescale entropy.See https://github.com/frisen-lab/TREX/issues/41#issuecomment-1757614961_