Supplement for:
Analysis of six types of protein-protein interfaces

Yanay Ofran ^1,2& Burkhard Rost ^{1, 2, 3, 4, *}

 

 

 

 

∆       Numbers give the percentage of how often the amino acid compositions of 1000 residues drawn at random from interface type X were most similar to different sets of 1000 residues drawn at random from one of the two interface types ('-' indicates values < 0.0). All values on the diagonal reflect the correct identification in our find-self test of the respective class. Off-diagonal elements reflect the confusion. The maximal standard deviation for each cell in the table is 4.4 percentage points.

 

 

	internal		external					no contact
	intra-domain	domain-domain		homo obligomer	homo complex	hetero obligomer	hetero-complex		no contact
intra-domain	73.3	20.2		0.7		5.6	0.2
domain-domain	17.5	63.5		1.3		16.7	1.0
homo obligomer	0.3	1.6		79.0		2.1	17.0
homo complex				0.1	99.9
hetero obligomer	3.6	17.5		1.7		72.0	5.2
hetero-complex		1.8		17.0		4.8	76.4
no contact									100.0

 

Same procedure as in Table 1, this time we included all the residues in our data sets that did not form contact of any type as a separate category

 

 

 

	internal		external					SWISS-PROT
	intra-domain	domain-domain		homo obligomer	homo complex	hetero obligomer	hetero-complex
intra-domain	73.8	20.1		0.7		3.9	0.1		1.4
domain-domain	17.3	62.3		1.5		17.6	1.1		0.2
homo obligomer	0.7	1.3		76.0		3.2	18.1		0.7
homo complex					100.0
hetero obligomer	3.4	16.9		3.1		71.5	4.3		0.8
hetero-complex		1.3		18.7		6.3	73.7
SWISS-PROT	2.2	0.5		1.8		2.3	0.1		93.1

 

Same procedure as in Table 1, this time we included all the residues in SWISS-PROT as a separate category.