Table I.

Benchmark and validation sets used for optimization and evaluation of PeptiMap peptide binding site identification

(A) Initial Calibration Benchmark (Extracted from PeptiDB 26; n=21)
Name	bound	unbound	Peptide sequence	Cath domain	domains/multimer	treatment	rank of hit	Pepsite218 ranks
dystrophin (WW)	1EG4A	1EG3A_1	NMTPYRSPPPYVPa	2.20.70.10 (b; 2x1.10.238.10)	1 of 4 different domains (47-84)	split: only first domain used	2,3	6,8,10
sh2a1 (SH2)	1D4TA	1D1ZA	KSLTIYAQVQK	3.30.505.10			1	-
lsb3 sla1 (SH3)	1SSHA	1OOTA	GPPPAMPARPT	2.30.30.40			2	1-6
erbb2 (PDZ)	1MFGA	2H3LA	EYLGLDVPV	2.30.42.10			1	-
wdr5 (WD40)	2H9MA	2H14A	ARTKQT	2.120.10.80	central hole not accessible to peptide	mask internal sites	3	1,3,5,6
usp7	2FOJA	2F1WA	GARAHSS	NA* (β-sandwich)			1,2	1-6
cyclophilin	1AWR	2ALFA	HAGPIA	2.40.100.10			1	1-10
p97 N-glycanase	2HPLA	2HPJA	DDLYG	NA* (p97)			1	-
traf2	1CZY	1CA4A	ace-PQQATDD	2.60.210.10	weak homotrimer (ABC)	use monomerc	3	-
i-ap1	1JD5A	1JD4A	AIAYFIPD	1.10.1170.10	weak homodimer (AB)	use monomer	1,2	4
gga1	1JWG_AC	1JWFA	DEDLLHI	1.25.40.90	weak homodimer (AC - in bound structure)	use monomer	2	-
ntf2	1GYB_AB	1GY7_AB	FSF	3.10.450.50	tight homodimer	unit: 2 chains	1 -	-
Clpx	2DS8_AB	2DS5_AB	ALRVVK	NA* (clpx)	tight homodimer, 2 sites	unit: 2 chains	1,2	-
calpain small subunit	1NX1A	1ALVAB	DAIDALSSDFT	1.10.238.10	tight homodimer central hole not accessible to peptide	unit: 2 chainsc	-	-
ap2	2VJ0A	1B9KA_1 1B9KA_2	PKGWVTFE FEDNFVP	2.60.40.1030 3.30.310.30	2 domains	split: both have peptide bound	- 3	- -
pim1 kinase transferase domain	2C3I	2J2IB_2	KRRRHPSG	(3.30.200.20) 1.10.510.10	2 domains	split: only second used	2	-
cdk2 cyclin	2CCHB	1H1RB	HTLKGRRLVFDN	1.10.472.10 x2	2 same domains	do not split	3	1,6
trypsin	2AGEX	1UTNA	Sin-AAPR	2.40.10.10 x2	2 same domains	do not split	1	6
Pcna	1RXZ	1RWZA	KSTQATLERWF	3.70.10.10	2 same domains	do not split	2	-
Endothiapepsin	1ER8	4APEA	PFHLLVY	2.40.70.10	2 same domains	do not split	1,2,3	1-6
(B) Validation set (Compiled from recent PDB releases 1-4/2013; n=9)
Name	bound	unbound	Peptide sequence	Cath domain	domains/multimer	treatment	rank of hit	Pepsite2 ranks
γ2 adaptin Ear domain	2YMT	4BCXA	EWGPWV	2.60.40.1230	1 domain		3	-
βcop WD40	4J73	3MKQA	EAKKLV	2x 2.130.10.10 + additional domain (based on 1VYH)	3 domains central hole not accessible to peptides	split, only first domain (2-300); site 1 masked	2,3	2,3,5,6
RADa	4B3B	4A74A	aceFHTA	3.40.50.300 (based on 1PZN)	weak homodimer	use monomer	2,3	-
FKBP35	4ITZA	3NI6A	sinALPFnit	3.10.50.40 (based on 1KT0A)	weak homodimer	use monomer	1	1-10
jnk1 kinase transferase domain	3VUH	3ELJ	PKRPTTLNLF	(3.30.200.20); 1.10.510.10	2 domains	split: only second domain	-	1-10
HIF alpha	4B7E	2W0X	EVVKLLLEHGADVLAQD	2-305; 1.10.287.1010	2 domains	split: only first domain	1,3	1,4,7-10
MLH1 mut alpha c-term	4FMNA	4E4WA	VRSKYFK	NA*	multidomain	too big without splittingc	- No split information available	-
demethylase	4FWF	4FWE	ARTMQTARKSTGG	(domain A02: 280-405 & 522-836, based on 2V1DA)	multidomain	split: only second domain	1	-
ck2 kinase	4IB5	3BQC	GCRLYGFKIHGCG	3.30.200.20; (1.10.510.10)	2 domains	split: only first domain	3	-

^aPeptide residues accurately mapped by fragments are highlighted in bold and underlined; Peptide residues with area covered by fragments are highlighted in bold but not underlined

^bXX: unassigned domains

^cBetter prediction obtained with SCOP classification (1CA4: 1,2 instead of 3), or domain parser (1ALV)

^*No domain assignment available based on CATH