ktp-forked-repos/tfmodisco

TF MOtif Discovery from Importance SCOres

TF-MoDISco

Installation: At the time of writing, the latest version on pypi is version 0.5.1.1 and can be installed using pip install modisco. To install from this source code, clone the repo and then run pip install --editable /path/to/cloned/repo.

Video of talk at NIPS MLCB: https://www.youtube.com/watch?v=fXPGVJg956E

Please see the following example notebooks:

• TF MoDISco TAL GATA: a self-contained example notebook that uses pre-computed importance scores (generated by a neural network) as input. Scores were generated using deeplift as illustated in this notebook
• TF MoDISco Nanog: a self-contained example notebook that uses pre-computed importance scores and an empirically-generated null distribution (generated by a gkm-SVM) as input. Scores were generated using gkmexplain as illustated in this notebook

A technical note describing version 0.4.2.2-alpha is available at https://arxiv.org/abs/1811.00416.

Full paper on the way.