Table 2.

Percentage of secondary structures affected by the preferential distribution

Dataset	Interface	Surface	Core	All
α-letters
Complete	2%	3%	4%	3%
Homodimers/Heterodimers	5%/3%	4%/5%	6%/6%	5%/5%
Obligate/Transient	9%/4%	4%/4%	4%/5%	4%/4%
Ubound/Bound	9%/9%	5%/2%	9%/6%	7%/4%
β-letters
Complete	10%	35%	17%	23%
Homodimers/Heterodimers	12%/11%	17%/12%	17%/17%	16%/14%
Obligate/Transient	10%/18%	20%/10%	18%/15%	18%/13%
Ubound/Bound	17%/17%	15%/14%	17%/18%	16%/16%
loop-letters
Complete	7%	10%	29%	14%
Homodimers/Heterodimers	8%/6%	9%/8%	29%/36%	13%/12%
Obligate/Transient	7%/11%	9%/8%	23%/30%	12%/13%
Ubound/Bound	12%/12%	11%/11%	38%/38%	17%/16%
border-letters
Complete	2%	5%	16%	7%
Homodimers/Heterodimers	3%/4%	6%/4%	17%/8%	8%/4%
Obligate/Transient	4%/10%	7%/3%	17%/6%	9%/5%
Ubound/Bound	12%/7%	5%/5%	13%/14%	7%/7%

The percentage of secondary structure affected by the preferential distribution is evaluated in the Interface, Surface and Core compartments using the difference between the observed number of the structural letters in a compartment and its expected number (given the repartition of the secondary structures type in the three compartments). The total effect (All) is evaluated according to the sum of the difference in the three compartments. The evaluation is performed on the seven datasets.