Analyzing Effects of Naturally Occurring Missense Mutations

Table 1

Methods	Short Summary	Examples (references)*	Some tools based on this method
Molecular dynamics	The trajectories of molecules are determined at atomic level by numerically solving the Newton’s equation of motion	(i) Thrombosis-related R2-FV haplotype: D2194G, Coagulation Factor V, domain C2 [8]; (ii) Parahemophilia, Factor V new brunswick: A221V, Coagulation Factor V, domain A [9]; (iii) FPLD, R482W; Lamin A/C [159]; (iv) Intellectual Disability: H101Q; CLIC2 protein [10]; (v) Snyder-Robin syndrome: G56S, V132G, I150T; spermine synthase; [5]	Eris [112, 132, 133] Tinker [158] GROMACS [160]
Molecular mechanics	Using molecular mechanics force field and optimization to model molecular systems	(i) 21-Hydroxylase-Deficiency: R132C, R149C, M283V, E431K; CYP450; C21 [161]; (ii) Cancer: A159V, A161V, N235I, N239Y, T256I, S269I; p53 [162]; (iii) Intellectual Disability: H101Q; CLIC2 protein; [10]; (iv) Mutability of human spermine synthase: all amino acids substitution at disease associated missense mutation sites G56, V132, and I150; human spermine synthase [6]; (v) Studying effects of nsSNPs on protein-protein interactions: nsSNPs in OMIM and non-OMIM; 264 protein-protein complexes with known nsSNPs located at the interface; [11]	FoldX [63, 64]
Monte Carlo simulation	Applying Monte Carlo sampling to predict preferred conformational states	(i) Noonan syndrome: D61Y, Tyrosine phosphatase SHP-2 [163]	IMC [164]
Electrostatic calculation	Calculating electrostatics energy and pKa/ionized states changes/electrostatic stability upon the missense mutations	(i) Snyder-Robinson Syndrome:; G56S, V132G, I150T human spermine synthase [5]; (ii) Thrombosis-related R2-FV haplotype: D2194G, Coagulation Factor V, domain C2 [8]; (iii) Noonan syndrome: D61Y, Tyrosine phosphatase SHP-2 [163]; (iv) Studying effects of nsSNPs on protein-protein interactions: nsSNPs in OMIM and non-OMIM; 264 protein-protein complexes with known nsSNPs located at the interface; [11]	DelPhi [165] MCCE [166–168] pKD [169]
Evolutionary properties	Based on structure and sequence analysis, for example, highly conserved residues in a protein family	(i) Homocystinuria: 204 mutations; cystathionine beta synthase; [170]	SNPs3D [138] PolyPhen [86]
Machine learning	learn the behavior of a system based on training datasets	(i) Snyder-Robinson Syndrome: G56S, V132G, I150T; human spermine synthase; [5]; (ii) Gastrointestinal stromal tumors: 19 mutations; KIT receptor [171]	I-Mutant 2.0/3.0 [71, 72, 134]
Graph methods	A branch of discrete mathematics. In protein science, this method is used to analyze the topological details of proteins with known structure	(i) Cancer: Y220C, R273H, R273C, R282W, and G245S; p53 protein; [172]; (ii) Predicting the structural effects of nsSNPs: 506 disease-associated nsSNPs; [173]	Bongo [173]
Statistical Potential	Based on the knowledge of statistical mechanics such as inverse Boltzmann law, ΔG = −kT log [gij(r)]	(i) Human X-linked Agammaglobulinemia (XLA): 16 missense mutations; Bruton’s tyrosine kinase (Btk); [174, 175]; (ii) Severe form of phenylketonuria: G46S; human phenylalanine hydroxylase (hPAH); [176]	DFIRE [55, 177, 178] PoPMuSiC-2.0 [179, 180] CUPSAT [181–183]
The BellKor collaborative filtering (CF) algorithm	Model relations of the known data points and the parameters of the model are learnt by the training database	(i) Using the known ΔΔG value to predict the ΔΔG value of other missense mutations at the same substitution site; 4803 mutants were used; [184]	Pro-Maya [184]