PlasForest: a homology-based random forest classifier for plasmid detection in genomic datasets

BMC Bioinformatics

Table 1 Compared performances of PlasForest and 4 other plasmid identification softwares on the testing dataset

Contig size	Index	PlasFlow	PlasClass	Mob-Recon	PPR-Meta	PlasForest
[50 bp, 1 kb)	Sensitivity	0.850	0.672	0.003	0.585	0.774
	Precision	0.373	0.473	1	0.550	0.972
	F1	0.518	0.555	0.006	0.567	0.862
	MCC	0.203	0.309	0.047	0.361	0.819
[1 kb, 2 kb)	Sensitivity	0.832	0.749	0.888	0.763	0.826
	Precision	0.294	0.451	0.973	0.558	0.986
	F1	0.434	0.563	0.928	0.644	0.899
	MCC	0.284	0.451	0.914	0.555	0.883
[2 kb, 5 kb)	Sensitivity	0.871	0.787	0.870	0.805	0.903
	Precision	0.308	0.464	0.946	0.552	0.975
	F1	0.455	0.584	0.907	0.655	0.938
	MCC	0.386	0.518	0.893	0.599	0.929
[5 kb, 50 kb)	Sensitivity	0.899	0.816	0.729	0.889	0.976
	Precision	0.366	0.432	0.932	0.413	0.971
	F1	0.520	0.565	0.818	0.564	0.973
	MCC	0.518	0.545	0.811	0.557	0.971
[50 kb, + ∞)	Sensitivity	0.912	0.761	0.500	0.944	0.999
	Precision	0.374	0.423	0.919	0.259	0.978
	F1	0.530	0.544	0.648	0.406	0.988
	MCC	0.559	0.543	0.669	0.459	0.988
Overall	Sensitivity	0.883	0.785	0.670	0.831	0.927
	Precision	0.348	0.443	0.939	0.431	0.973
	F1	0.499	0.566	0.782	0.568	0.950
	MCC	0.483	0.531	0.775	0.540	0.945

ISSN: 1471-2105