A novel strategy for classifying the output from an in silicovaccine discovery pipeline for eukaryotic pathogens using machine learning algorithms

Table 3 Sensitivity and specificity performance measures of binary classification for individual input variables taken from datasets

			Datasets (comprising evidence profiles)
Input variable^a	Type^b	Data^c	T. gondii		Plasmodium		C. elegans		Benchmark
			SN	SP	SN	SP	SN	SP	SN	SP
Phobius_TM	TM	D	0.57	0.89	0.85	0.90	0.91	0.97	0.74	0.93
Phobius_SP	SP	T	0.52	0.89	0.39	1.00	0.25	0.99	0.49	0.96
SignalP	SP	C	0.52	1.00	0.39	1.00	0.25	1.00	0.39	1.00
TargetP_SP	SP	C	0.67	1.00	0.77	1.00	0.34	1.00	0.56	1.00
TargetP_loc	SP	T	0.67	0.94	0.76	1.00	0.27	1.00	0.56	1.00
TMHMM_AA	TM	C	0.62	0.89	0.66	0.98	0.91	1.00	0.80	1.00
TMHMM_First60	SP	C	0.43	0.93	0.26	1.00	0.37	1.00	0.49	0.97
TMHMM_TM	TM	D	0.57	0.89	0.65	1.00	0.90	1.00	0.77	1.00
WoLF_PSORT	Sub	C	0.76	0.94	0.42	1.00	0.77	0.98	0.60	0.97
WoLF_PSORT_annotation	Sub	T	1.00	0.56	0.92	0.74	1.00	0.72	0.96	0.73
MHCI	B	C	0.76	0.56	0.78	0.84	0.77	0.69	0.74	0.84
MHCII	B	C	0.86	0.39	0.80	0.74	0.90	0.52	0.54	0.84

Abbreviations: SN = sensitivity; SP = specificity; T. gondii = Toxoplasma gondii; Plasmodium = species in the genus Plasmodium including falciparum, yoelii yoelii, and berghei; C. elegans = Caenorhabditis elegans; Benchmark = dataset comprising evidence for T. gondii and Neospora caninum proteins from published studies.
^aInput variable = predicted protein characteristic (i.e. a column from evidence profile).
^bType = prediction type: transmembrane domains (TM), secretory signal peptide (SP), sub-cellular location (Sub), peptide-MHC binding (B).
^cData = data type: discrete (D), continuous (C), text (T).
The values underlined denote the best performing input variable for classifying the published proteins.
Test criteria on input variable for binary classification:
Phobius_TM: YES if number of transmembrane domains > 0 else NO.
Phobius_SP: YES if = ‘Y’ else NO.
SignalP: YES if > 0.5 else NO.
TargetP_SP: YES if > 0.5 else NO.
TargetP_loc: YES if = ‘S’ else NO.
TMHMM_AA: YES if > 0 18^$$ else NO.
TMHMM_ First60: YES if > 10^$$ else NO.
TMHMM _TM: YES if number of transmembrane domains > 0 else NO.
Wolf_PSORT: YES if > 16^$$ else NO.
WoLF_PSORT_annotation: YES if = ‘membrane’ or ‘secreted’ else NO.
MHCI: YES if > 0.5 else NO.
MHCII: YES if > 0.5 else NO.
^$$A value recommended by the creator of the program. $Specificity = \frac{True Negatives}{True Negatives + False Positives}$ $Sensitivity = \frac{True Positives}{True Positives + False Negatives}$

ISSN: 1471-2105