Single nucleotide polymorphism genes and mitochondrial DNA haplogroups as biomarkers for early prediction of knee osteoarthritis structural progressors: use of supervised machine learning classifiers

Table 4 Synergy analysis in PVBSP forecasting

Scenario 1
Model	Input combinations	Accuracy (%)
Model	Input combinations	Train	Test
Model 1	Equation 1 (12 variables)	83.2	88.8
M31	Equation 2 (5 variables)	75.5	78.8
M32-1	M31 + GNL3	81.4	82.7
M32-2	M31 + MCF2L	77.0	81.3
M32-3	M31 + FTO	82.7	85.0
Scenario 2
No. of inputs	Model no	Accuracy (%)
No. of inputs	Model no	Train	Test
(3 + 1) variables	MH2	75.5	80.0
Age, BMI, mtDNA haplogroup, FTO
(3 + 2) variables	MH17	85.2	82.5
Age, BMI, mtDNA haplogroup, FTO, SUPT3H
(3 + 3) variables	MH46	82.1	88.8
Age, BMI, mtDNA haplogroup, FTO, SUPT3H, GNL3
(3 + 4) variables	MH80	82.7	88.8
Age, BMI, mtDNA haplogroup, TP63, DUS4L, GDF5, TGFA
(3 + 5) variables	MH101	83.2	88.8
Age, BMI, mtDNA haplogroup, TP63, GNL3, DUS4L, GDF5, TGFA
(3 + 6) variables	MH106	83.2	88.8
Age, BMI, mtDNA haplogroup, TP63, FTO, SUPT3H, GNL3, DUS4L, GDF5

Model 1 is Eq. 1: PVBSP = f(age, BMI, mtDNA haplogroup, cluster, TP63, FTO, GNL3, DUS4L, GDF5, SUPT3H, MC2FL, TGFA), and M31, Eq. 2: PVBSP = f(age, BMI, TP63, DUS4L, GDF5)
M Model, No. Number, PVBSP Probability values of being structural progressors, test Testing stage, train Training stage

ISSN: 1741-7015