A Novel Approach Utilizing Domain Adversarial Neural Networks for the Detection and Classification of Selective Sweeps.

We considered two possibilities: (i) DASDC corrects potential differences between simulated and real genomic features, thereby reducing the impact of distributional differences and obtaining more accurate classification results, and (ii) there may be significant differences in characteristics between the simulated genome and the real genomes, with LCT genes undergoing strong natural selection that could exhibit features similar to those observed during a hard sweep in simulated genome data.

also in 13,097 other papers