Predicting functional sites with an automated algorithm suitable for heterogeneous datasets

BMC Bioinformatics

Table 3 Structural assessment of the phylogenetic motif functional site predictions^a

Protein family	PDB ID	Correct	Useful info	Wrong
A. Proteins from Table 1
Acetyglucosamine-6-phoshate deacyetylase^b	1O12	6	1	0
Alcohol dehydrogenase	1JVB	2	1	1
CuZnSOD	1SPD	3	2	1
Cytochrome P450	1N4G	6	4	0
Enolase	2ONE	6	2	0
Glyceroladhyde-3-phosphate dehydrogenase	1DC4	2	2	3
Glycerol kinase	1B05	7	1	0
Glutamate dehydrogenase	1HWZ	7	0	0
Inorganic pyrophosphatase	1I6T	3	2	0
Myoglobin	1MBA	4	1	0
Succinate dehydrogenase – FAD	1NEK	4	1	2
Succinate dehydrogenase – Fe/S	1NEK	4	0	0
Succinate dehydrogenase – Heme	1NEK	5	0	0
TATA box binding protein	1TBP	n/a^c
Triosephosphate isomerase	7TIM	3	4	1
B. Additional examples
Acetate kinase	1G99	4	4	0
Aconitase A	7ACN	8	3	5
Alanine racemase	1L6F	4	1	1
Arginyl-tRNA synthetase	1F7U	2	2	1
Biotin carboxylase^b	1DV2	3	3	2
Catalase	1GGF	9	0	0
Citrate synthase	1NXG	8	1	0
Isocitrate dehydrogenases	1IA2	3	0	0
Malate/lactate dehydrogenases	1EMD	6	1	0
Malate synthase	1P7T	6	0	0
Phosphomannomutase	1P5D	5	2	0
Selenocysteine lyase	1ECX	5	2	1
Threonine aldolase	1LW5	6	2	0
Thymidylate synthase	1AIQ	4	0	1
Transaldolase	1ONR	1	1	1
Trehalose-6-phosphate synthase	1UQU	5	3	3

^a A sequence window width of 5 is used for each of the above examples and gap range of {-1: -2}.
^b The known functional site defined in Acetyglucosamine-6-phoshate deacyetylase and biotin carboxylase is prohibitively incomplete; as such, functional sites indicated in [41] and [42], respectively, are also included. ^c Because so much of the TATA-box binding protein surface area directly interacts with its DNA substrate, the known functional site sphere encompasses nearly the entire protein, making the assessment unsporting because all predictions will trivially be correct.

ISSN: 1471-2105