DV-TrioTrain v0.8
DV-TrioTrain is a model development framework for DeepVariant: a deep-learning-based germline variant caller. Here we provide an automated pipeline for extending DeepVariant v1.4.0 in new species.
Models built by DV-TrioTrain do not require trio-binned data for variant calling.
Our unique re-training approach enables the model to incorporate inheritance expectations. Using TrioTrain to build new versions of DeepVariant models requires trio-binned genomes, the final model produced by TrioTrain works on individual genomes, while prioritizing features of inherited variants.
Contributing to TrioTrain
Please open a Github pull request if you wish to contribute to TrioTrain.
License
Last update:
March 8, 2024