#ML based #Protein-Protein Interaction prediction models report very high accuracy metrics but can be very inaccurate in practice. In this opinion paper, we discuss (and empirically show) that a bias in the use of training data is to blame. Read more: htt
RT @FrontBioinf: New Research: On the choice of negative examples for prediction of host-pathogen protein interactions https://t.co/GYnYWln…
RT @FrontBioinf: New Research: On the choice of negative examples for prediction of host-pathogen protein interactions https://t.co/GYnYWln…
RT @FrontBioinf: New Research: On the choice of negative examples for prediction of host-pathogen protein interactions https://t.co/GYnYWln…