BlueSTARR v0.1.1 is the updated version that was trained on a full-set dataset containing sequences from all chromosomes, unlike the original model (v0.1.0), which excluded chromosomes 1 and 2 for testing and validation. It uses DNA and RNA counts normalized by library size, employs unbiased down-sampling for training, and optimizes predictions using Mean Squared Error (MSE) as the loss function
bill-majoros:bluestarr-k562-model-v0.1.1