PEER: A Comprehensive and Multi-Task Benchmark for Protein Sequence Understanding (Supplementary Material) Jiarui Lu
–Neural Information Processing Systems
The DDE protein sequence feature vector is defined by the statistical features of dipeptides, i.e., two consecutive amino acids in the protein sequence. For example, the feature of dipeptide "st" is defined by its dipeptide composition (D The Moran feature descriptor defines the distribution of amino acid properties along a protein sequence. The Moran feature vector is with 8M dimensions (M is the parameter of maximum lag, setting as 30 following iFeature). Table 1: Balanced metric (weighted F1) compared with accuracy on multi-class classification tasks. We report mean (std) for each experiment.
Neural Information Processing Systems
Feb-10-2025, 17:43:52 GMT