Views
No views yet
20211113_021200), converted to
safetensors from gs://tewhey-public-data/CODA_resources/ with no retraining or
modification.1from modeling_malinois import MalinoisModel
2
3model = MalinoisModel.from_pretrained("saarantras1/malinois").eval()
4
5preds = model.predict(["ACGT" * 50]) # (n, 3): K562, HepG2, SKNSHpredict rather than calling the model directly: it adds the MPRA vector
context the model was trained with (a bare 200mer is not valid input) and averages
over both strands. Skipping either step returns plausible-looking but wrong numbers
instead of an error.predict reverse-complements the 200 bp insert and re-flanks it in
the forward orientation, following src/vcf_predict.py in the upstream code base.
The CODA tutorial notebook instead flips the assembled 600 bp construct. Both appear in upstream
code; this repository uses the former.1@article{gosai2024coda,
2 title = {Machine-guided design of cell-type-targeting cis-regulatory elements},
3 author = {Gosai, Sager J. and Castro, Rodrigo I. and Fuentes, Natalia and
4 Butts, John C. and Mouri, Kousuke and Alasoadura, Michael and
5 Kales, Susan and Nguyen, Thanh Thanh L. and Noche, Ramil R. and
6 Rao, Arya S. and Joy, Mary T. and Sabeti, Pardis C. and
7 Reilly, Steven K. and Tewhey, Ryan},
8 journal = {Nature},
9 year = {2024},
10 doi = {10.1038/s41586-024-08070-z}
11}