Review Article

A Survey of Computational Intelligence Techniques in Protein Function Prediction

Table 2

Summary of computational intelligence (CI) techniques in prediction of subcellular localization.

ReferenceCI techniquesPerformanceDatasets

[47]SVMAccuracy: 86.3%Amino acid compositions
[48]SVMAverage accuracy: 66.7%Functional domain composition of protein
[49]SVMOverall recall: 89.8%Amino acid subsequence
[50]SVMOverall accuracy: 93.1%Physiochemical property of amino acid
[51]SVMAccuracy: 84.9%Amino acid composition, dipeptide composition, and similarity information
[52]SVMAccuracy: 91.2%Compositions of residues, dipeptides, and physicochemical properties
[53]-NNOverall accuracy: 80%Dipeptide composition of amino acids
[54]-NNOverall accuracy: 92.5%Amino acid compositions, dipeptide compositions, and physicochemical properties
[55]-NNOverall accuracy: 85.4%Functional domain composition
[56]-NNOverall accuracy: 93.57%PSSM and pseudoamino acid composition
[57]SVMOverall accuracy: 74.00%N-terminal targeting sequences amino acid composition and protein sequence motifs
[58]CSVMOverall accuracy: 80.03%Pseudoamino acid composition
[59]SVM with GAOverall accuracy: 72.82%Physiochemical property of amino acid
[60]SVMOverall accuracy: 90.96% and MCC: 0.8655Combination of sequence alignment and feature based on amino acid composition
[61]SVMAccuracy: 73.71%Amino acid composition and PSSM
[62]SVMAccuracy: 88.3%Pseudoamino acid composition
[63]SVMRecall: 91.30%Sequence motifs
[64]SVMAccuracy: up to 94.00%Amino acid composition, amino acid pair, 1, 2.3 gapped amino acid pair compositions
[65]SVMAccuracy: up to 93%Integrates features from phylogenetic profiles and gene ontology
[66]Recurrent NNOverall accuracy: 72.55% Pseudo amino acid composition
[67]N-to-1 NNAccuracy: up to 89%Protein sequence
[68]SVMOverall accuracy: 93.57%Amino acid and dipeptide, composition, reduced physiochemical properties, gene ontology, PSSM, and pseudoaverage chemical shift
[69]SVM and ANNAccuracy: 68%Structural properties of a protein