Skip to main content
Fig. 3 | BMC Medical Genomics

Fig. 3

From: NeoMutate: an ensemble machine learning framework for the prediction of somatic mutations in cancer

Fig. 3

Comprehensive performance evaluation of different approaches. Only those approaches having a sensible recall (> 0.5) were chosen for the comparison. a) Individual variant callers raw results evaluation. b) Standard filtering results evaluation. m2 s2: mutect2 and strelka2 calls intersection; m2s2_HQ: mutect2 and strelka2 HQ (only variants tagged as `PASS`) calls intersection; cons_n: consensus voting (intersection) of at least n tools; cons_2_HQ: consensus voting of the HQ call sets of least 2 tools. b) ML results evaluation. LRC: logistic regression classifier; SVMl: support vector machine classifier with linear kernel; DT: decision tree; GNB: Gaussian Naïve Bayes; RFC: random forest classifier; GBDT: gradient boosting decision tree; NN: neural network

Back to article page