--The benchmark dataset:
--The extracted features
a: HHSearch is contained in HHsuite ;
b: Top-1-gram feature is generated by Pse-in-one2.0 ;
c: Top-2-grams feature is generated by Pse-in-one2.0 ;
d: DeepFR is the result of this study ;
e: 84-features is the result of this study .
--The features for LTR training
a: The processing of features refers to Fold-LTR-TCP .
--The trained Learning to Rank model
1. Liu B, Zhu Y, Yan K. Fold-LTR-TCP: protein fold recognition based on triadic closure principle, Brief Bioinform 2019.
2. Lindahl E, Elofsson A. Identification of related proteins on family, superfamily and fold level, J Mol Biol 2000;295:613-625.
3. Liu B, Wu H, Chou K-C. Pse-in-One 2.0: An Improved Package of Web Servers for Generating Various Modes of Pseudo Components of DNA, RNA, and Protein Sequences, Natural Science 2017;Vol.09No.04:23
4. Remmert M, Biegert A, Hauser A et al. HHblits: lightning-fast iterative protein sequence searching by HMM-HMM alignment, Nat Methods 2011;9:173-175.
5. Zhu J, Zhang H, Li SC et al. Improving protein fold recognition by extracting fold-specific features from predicted residue–residue contacts, Bioinformatics 2017;33:3749-3757.
6. Jo T, Cheng J. Improving protein fold recognition by random forest, BMC Bioinformatics 2014;15 Suppl 11:S14.
7. Burges CJ. From ranknet to lambdarank to lambdamart: An overview, Learning 2010;11:81.