Data and reproducibility - AngeloDAngelo/Ktools GitHub Wiki
Data and reproducibility
Included examples
- Synthetic demo: 24 reference and 48 control sequences, generated with seed 42. The demo creates fictional RBP scores using seed 123. Inputs and demo outputs.
- CDR1as: a public circBase-derived sequence mapped to GRCh38, with seed annotations and example KRS scores. The comparison FASTA is supplied by the user. Sequence source and launch command.
Reference resources
- PEKA matrices: choose one of the two included score matrices.
- miRNA family tables: complete table and human subset; the complete table is also installed with the Python package.
- UniProt BED inputs: required file formats for K-CDS. Provide BED tracks and a matching GTF before running this module.
Run records
Record the code commit, input checksums, annotation release, group membership, parameters and random seeds. Preserve full result tables alongside selected signatures. Save dependency versions with:
python -m pip freeze > environment.txt