MSLP: mRNA subcellular localization predictor based on machine learning techniques

BMC Bioinformatics

Table 7 Performance of MSLP and other methods on cross validation for five locations

Location	Metric	mRNALoc	mLoc-mRNA	iLoc-mRNA	MSLP
Cytoplasm	Sn	66.69	73.24	Com	84.60
	Sp	67.41	68.51	Com	81.60
	Acc	67.10	70.87	Com	83.10
Nucleus	Sn	74.17	72.89	Com	86.80
	Sp	73.22	73.99	Com	86.00
	Acc	73.59	73.44	Com	86.40
ER (C3 for iLoc-mRNA)	Sn	74.09	63.04	89.72	80.80
	Sp	75.49	73.68	97.56	78.70
	Acc	75.36	68.36	NR	79.70
ExR	Sn	62.67	–	–	70.10
	Sp	65.34	–	–	70.20
	Acc	65.20	–	–	70.10
Mitochondria	Sn	96.28	98.53	C	97.80
	Sp	99.80	91.70	Com	98.10
	Acc	99.70	96.46	Com	97.90

For our method MSLP, we mentioned the results from the CatBoost model.–: Location was not considered in the literature. Com: Location was combined with other locations, therefore not comparable; NR: Not reported. Bold numbers highlight the best value of performance metrics

ISSN: 1471-2105