Dissecting protein loops with a statistical scalpel suggests a functional implication of some structural motifs

BMC Bioinformatics

Table 5 Correspondence between extreme superfamily-specific words and Swiss-Prot annotations in the initial data set

			Statistics in the initial dataset		Comparison with Swiss-Prot
Word	Occ ^a	Lp _max	nb_sf*/ nb_sf^b	Superfamilies ^c	Annot	Match/total (Precision (%)) ^d	Sensitivity (%)
URNH	43	54.95	1/17	48726*	Disulfide	7/14 (50)	4
RNHB	59	51.33	1/28	48726*	Disulfide	9/20 (45)	6
UQHS	53	75.07	1/16	52058*	Repeat	12/22 (55)	41
SUQH	70	63.42	1/25	52058*	Repeat	11/26 (42)	38
QHSG	37	51.75	1/12	52058*	Repeat	4/10 (40)	14
HSGI	63	76.26	1/18	52058*	Repeat	5/12 (42)	17
QXUS	43	52.05	1/10	51735*	Repeat	1/15 (7)
ZSGI	99	52.22	1/49	52058*	Repeat	7/36 (19)
GSUS	169	140.49	3/59	141571*, 52047, 52058	Repeat	6/38 (16)
GZDO	115	84.72	3/49	47473*, 52833, 52935	Repeat	1/35 (3)
DODQ	73	157.01	1/17	47473*	CA_BIND	15/23 (65)	75
ZDOD	48	91.27	1/13	47473*	CA_BIND	11/16 (69)	58
YUOD	111	184.67	1/11	52540*	NP_BIND	39/41(95)	35
UODO	142	210.14	4/14	52540*,53659, 54211, 55729	NP_BIND	49/60 (82)	38
OEIJ	33	53.84	1/4	51735*	NP_BIND	6/7 (86)	14
EIJU	48	51.68	1/13	51735*	NP_BIND	7/15 (47)	20
USLG	121	137.35	2/47	141571*, 51206	NP_BIND	2/22 (9)
UZCI	99	63.70	2/28	103025*, 56784	NP_BIND	1/13 (8)
RUDO	27	55.55	1/4	53335*	Binding	5/10 (50)	18
UGRU	37	60.07	1/8	53335*	Binding	4/12 (33)
EGZD	48	51.68	1/5	51735*
GRUD	33	70.55	1/6	53335*
SLGS	60	118.45	1/17	141571*

This comparison is made on a subset on the initial set: 1487 proteins that can be mapped to biological annotations using the PDB/UniProt Mapping database. ^a: word occurrence. ^b: nb_sf denotes the number of SCOP superfamilies in which the structural word is seen. ^c: superfamilies in which the word is over-represented. ^d: match and total denote the number of fragments annotated and the total number of fragments, respectively. Bold font indicates a match/total ratio greater than 40%. Italic font indicates a match/total ratio lower than 40%. Abbreviations used: NP BIND = nucleotide phosphate-binding site, CA_BIND = calcium-binding site. SCOP ids: 103025 = Folate-binding domain, 141571 = Pentapeptide repeat-like, 47473 = EF-hand, 48726 = Immunoglobulin, 51206 = cAMP-binding domain-like, 51735 = NAD(P)-binding Rossmann-fold domains, 52047 = RNI-like, 52058 = L domain-like, 52540 = P-loop-containing nucleoside triphosphate hydrolases, 52833 = Thioredoxin-like, 52935 = PK C-terminal domain-like, 53335 = S-adenosyl-L-methionine-dependent methyltransferases, 53659 = Isocitrate/isopropylmalate dehydrogenase-like, 54211 = Ribosomal protein S5 domain 2-like, 55729 = Acyl-CoA N-acyltransferases (Nat), 56784 = HAD-like. "*" denotes the superfamily in which the word is most over-represented.

ISSN: 1471-2105