SDNN-PPI: self-attention with deep neural network effect on protein-protein interaction prediction
Problem
Identifying protein-protein interactions by wet-lab experiments is costly and time-consuming.
Why it matters
PPIs govern transcriptional regulation, signal transduction and drug signalling, so cheaper prediction aids biology and drug research.
What the AI did
SDNN-PPI uses a deep neural network with self-attention over AAC, CT and AC sequence features to predict PPIs computationally.
Results
5-fold cross-validation accuracy 95.48% (S. cerevisiae core subset) and 98.94% (human); 93.15% and 88.33% on interspecific datasets.
Impact
Cross-validated accuracies suggest a faster computational alternative to wet-lab PPI detection.
Record
| Venue | BMC Genomics |
|---|---|
| Published | 2022 |
| Work type | article |
| Authors | Xue Li, Peifu Han, Gan Wang, Wenqi Chen, Shuang Wang, Tao Song |
| Citations | 83 (OpenAlex, at import time) |
| References | 64 |
| Model or method | SDNN-PPI (self-attention DNN) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1186/s12864-022-08687-2 |
| Open access | gold |
| Keywords | protein-protein interaction, deep learning, self-attention, sequence features, cross-validation |
Importance rating
significant — thousands to millions affected
Abstract states PPIs govern transcriptional regulation, signal transduction, drug signalling; wet-lab detection is costly and time-consuming; no scale.