| _version_ | 1866901715784040448 |
|---|---|
| author | Crnjar, Alessandro Koo, Peter Desmarais, John |
| author_facet | Crnjar, Alessandro Koo, Peter Desmarais, John |
| contents | <p>Data for reproducibility for the paper: "PIONEER: a virtual platform for iterative improvement of genomic deep learning".</p> <p>Systems:</p> <ul> <li>LentiMPRA (K562) </li> <li>STARR-seq (DeepSTARR)</li> </ul> <p>Main Experiments per system:</p> <ul> <li>Sequence generation: UGM, Mutagenesis, Random, All.</li> <li>Seuqnece Selection: Random, Uncertainty, LCMD.</li> </ul> <p>Supplementary Information (SI) Experiments:</p> <ul> <li>Fair Cost Mutagenesis, Random, All.</li> <li>10% and 25% mutation ratio: UGM, Mutagenesis.</li> <li>UGM with Single Oracle.</li> <li>UGM with Deep Ensemble.</li> </ul> <p>Data directories:</p> <ul> <li>data_LentiMPRA/ : proposed sequences and labels per cycle, LentiMPRA.</li> <li>data_STARR-seq/: proposed sequences and labels per cycle, STARR-seq.</li> <li>covariate_test_shifts: test sequences and labels corresponding to no shift, small shift, large shift (high activity), large shift (low activity).</li> <li>weights_models/: model weights per cycle and oracle weights.</li> <li>oracles_train_data/: sequences and labels for the training of the oracles.</li> </ul> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_15045788 |
| institution | Zenodo |
| language | |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Data for: "PIONEER: a virtual platform for iterative improvement of genomic deep learning" Crnjar, Alessandro Koo, Peter Desmarais, John <p>Data for reproducibility for the paper: "PIONEER: a virtual platform for iterative improvement of genomic deep learning".</p> <p>Systems:</p> <ul> <li>LentiMPRA (K562) </li> <li>STARR-seq (DeepSTARR)</li> </ul> <p>Main Experiments per system:</p> <ul> <li>Sequence generation: UGM, Mutagenesis, Random, All.</li> <li>Seuqnece Selection: Random, Uncertainty, LCMD.</li> </ul> <p>Supplementary Information (SI) Experiments:</p> <ul> <li>Fair Cost Mutagenesis, Random, All.</li> <li>10% and 25% mutation ratio: UGM, Mutagenesis.</li> <li>UGM with Single Oracle.</li> <li>UGM with Deep Ensemble.</li> </ul> <p>Data directories:</p> <ul> <li>data_LentiMPRA/ : proposed sequences and labels per cycle, LentiMPRA.</li> <li>data_STARR-seq/: proposed sequences and labels per cycle, STARR-seq.</li> <li>covariate_test_shifts: test sequences and labels corresponding to no shift, small shift, large shift (high activity), large shift (low activity).</li> <li>weights_models/: model weights per cycle and oracle weights.</li> <li>oracles_train_data/: sequences and labels for the training of the oracles.</li> </ul> |
| title | Data for: "PIONEER: a virtual platform for iterative improvement of genomic deep learning" |
| url | https://doi.org/10.5281/zenodo.15045788 |