Spliceator: multi-species splice site prediction using convolutional neural networks

Scalzitti, Nicolas; Kress, Arnaud; Orhand, Romain; Weber, Thomas; Moulinier, Luc; Jeannin-Girardon, Anne; Collet, Pierre; Poch, Olivier; Thompson, Julie D.

doi:10.1186/s12859-021-04471-3

BMC Bioinformatics

Table 2 Performance of Spliceator and state-of-the-art programs on six independent benchmarks

From: Spliceator: multi-species splice site prediction using convolutional neural networks

	Accuracy		Precision		Sensitivity		Specificity		F1 Score
	Donor	Acceptor	Donor	Acceptor	Donor	Acceptor	Donor	Acceptor	Donor	Acceptor
Human
Spliceator	92.7	89.5	90.0	86.1	96.1	94.4	89.4	84.7	93.0	90.0
SpliceFinder	91.5	91.1	76.9	73.3	97.0	99.8	89.6	88.2	85.8	84.5
DSSP	93.2	91.0	96.9	94.4	89.2	87.1	97.2	94.8	92.9	90.6
MaxEntScan	91.8	83.7	88.8	77.3	95.8	95.6	87.9	71.9	92.1	85.5
NNSplice	68.4	65.9	62.3	61.5	89.2	86.7	48.5	44.8	73.4	72.0
Fish
Spliceator	95.0	91.3	91.6	86.0	99.0	98.7	90.9	84.0	95.1	91.9
SpliceFinder	92.8	94.9	81.5	85.0	96.4	99.8	91.4	93.0	88.3	91.8
DSSP	94.6	93.6	97.5	94.2	91.5	92.9	97.6	94.3	94.4	93.6
MaxEntScan	93.9	83.3	90.6	75.4	98.1	99.0	89.8	67.7	94.2	85.6
NNSplice	71.0	68.9	64.3	63.9	90.8	85.6	52.2	52.5	75.3	73.2
Fly
Spliceator	94.6	90.4	91.6	86.2	98.3	96.4	91.0	84.5	94.8	91.0
SpliceFinder	92.1	90.9	79.8	72.8	95.8	99.8	90.7	88.0	87.1	84.2
DSSP	94.0	91.6	96.6	93.1	91.3	89.8	96.7	93.4	93.9	91.4
MaxEntScan	94.5	86.0	91.8	79.5	97.6	97.0	91.3	75.0	94.6	87.4
NNSplice	71.7	71.1	65.3	64.8	91.4	92.1	52.3	50.3	76.2	76.1
Worm
Spliceator	93.9	88.8	91.4	86.9	96.9	91.4	90.9	86.2	94.1	89.1
SpliceFinder	88.7	87.9	71.2	64.1	93.3	99.3	87.1	84.8	80.8	77.9
DSSP	90.4	84.1	97.0	93.1	83.4	73.6	97.4	94.5	89.7	82.2
MaxEntScan	92.9	82.0	91.6	76.8	94.5	92.0	91.3	72.1	93.0	83.7
NNSplice	65.8	57.8	61.9	58.0	80.3	56.4	51.7	59.2	69.9	57.1
Plant
Spliceator	94.7	90.6	91.8	89.1	98.1	92.7	91.3	88.6	94.9	90.8
SpliceFinder	88.9	87.6	72.2	63.1	92.9	99.3	87.5	84.4	81.3	77.2
DSSP	88.2	85.1	96.5	94.2	79.3	74.8	97.1	95.4	87.1	83.3
MaxEntScan	92.4	85.3	90.5	79.2	94.6	95.9	90.1	74.8	92.5	86.7
NNSplice	59.8	59.2	57.2	57.1	78.4	71.3	41.0	47.3	66.2	63.4
PVP
Spliceator	86.0	83.5	82.9	83.1	90.8	84.0	81.2	82.9	86.6	83.6
SpliceFinder	87.3	89.3	67.9	58.8	77.6	98.6	89.9	87.7	72.4	73.8
DSSP	84.3	76.6	92.5	84.7	74.6	65.0	93.9	88.2	82.6	73.5
MaxEntScan	84.5	80.0	81.0	75.4	90.1	89.4	79.1	70.4	85.2	81.8
NNSplice	64.8	63.7	60.4	61.4	85.8	73.9	43.6	53.5	70.9	67.1
Average
Spliceator	92.82	89.02	89.88	86.23	96.53	92.93	89.12	85.15	93.08	89.40
SpliceFinder	90.22	90.28	74.92	69.52	92.17	99.43	89.37	87.68	82.62	81.57
DSSP	90.78	87.00	96.17	92.28	84.88	80.51	96.65	93.43	90.10	85.77
MaxEntScan	91.67	83.38	89.05	77.27	95.12	94.82	88.25	71.98	91.93	85.12
NNSplice	66.92	64.43	61.90	61.12	85.98	77.67	48.22	51.27	71.98	68.15

Performance metrics for Spliceator, SpliceFinder, DSSP, MaxEntScan, and NNSplice, using six independent benchmarks from model and non-model organisms: human, fish, fly, worm, plant and PVP (Protists + Viridiplantae) (values in bold correspond to the highest performance)

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com