Neural predictors do not give more accurate gene-pair results than a sum
Large differences from the sum show large errors.
Claimed, not confirmed
Researchers compared 8 methods on 3 Perturb-seq datasets to find the effects of gene-pairs with no measured effect. Neural predictors did not give more accurate results than the sum of the 2 single-gene effects. The difference from that sum is additive disagreement and is a good signal of error in all 3 datasets.
Sources
Posted
