SCALA: Sequence Clustering Against Leaking informAtion

A script constructing the most challenging training-test-validation dataset by hierarchically clustering a database of fasta sequencing and then separating the tree such that no previously seen similar sequence is in the test or validation set.

Usage

python3 scala.py -i <path_to_fasta_database> -o <directory_for_outputfiles>

Additional optional parameters:
-s : clustering steps (int) (default=4)
-f : additional fasta output (boolean flag)
-tr : size of training set (default=60)
-te : size of test set (default=30)\

Output: