Deep learning for variant calling
![](/project/ml/featured_huf9954dcf5270275cb357acbdb41b9f25_25865_720x0_resize_lanczos_2.png)
Experimenting with new ways of generating pileup images from DNA sequencing data to increase variant calling accuracy. The machine learning pipeline involves parsing sequencing data files and creating pileup images, as well as training CNNs for classification (written with PyTorch and fastai).