Fixed bugs related to overexpressing genes

#229
by davidjwen - opened
  • indexing was off by one with overexpressing "all"
  • perturbed genes should be removed correctly before calculating the cosine similarities
  • code cleaned code cleaned up slightly

known bug: the embeddings are very slightly different (< 1% cosine shift) if a cell is calculated on its own vs. in a batch

ctheodoris changed pull request status to merged

Sign up or log in to comment