. Author manuscript; available in PMC: 2011 Aug 1.

Published in final edited form as: Proteins. 2010 Aug 1;78(10):2338–2348. doi: 10.1002/prot.22746

Table I.

Comparison Between the Results Given by Wild-Type Sequences, Sequences Designed by RosettaDesign 2.3, RosettaDesign and Structure-Derived Sequence Profile (RosettaDesign-S), RosettaDesign-S Plus Repetitive, and Reference-State Energy Terms (RosettaDesign-SR)

	% Polar residues^a
	Seq. ID (%)^b	% LC^c	% LLL^d	% VVV^e	All	α	β	Coil
Wild-type	100	2.1	0.030	0.060	49.1	51.2	41.4	52.6
RosettaDesign	31.9	6.4	0.113	0.033	44.0	48.4	33.8	46.9
Sequence profile^f	24.2	15.1	0.030	0.714	28.2	24.2	21.0	38.3
RosettaDesign-S	35.6	21.3	0.783	0.597	37.4	36.9	28.2	45.1
RosettaDesign-SR	34.4	1.4	0.029	0.030	48.2	51.3	40.0	50.9

The results are obtained for the training set of 33 proteins averaged over 100 sequences designed for each protein.

Frequency of polar residues for helical, strand, and coil positions, respectively. Polar residues include D,E,H,K,N,Q,R,S,T, and Y.

The average sequence identity to wild type sequences of target proteins.

Percent of low-complexity regions defined by program SEG⁷².

Percent of occurrence for three sequentially linked Leu residues.

Percent of occurrence for three sequentially linked Val residues.

Results based on the consensus sequence from structure derived sequence profile.