Optimization algorithms for functional deimmunization of therapeutic proteins

BMC Bioinformatics

Table 3 Full-length Epo.

Variant	Epo 1-166	E	B	C	ΔΔG°	S
wild type	Epo	136	0	0.00	0.00	74
1 substitution
BLOSUM	R110Q	126	0	2.87	0.70	72
Conservation	R110G	128	1	2.81	1.87	75
FoldX	N147D	124	1	2.22	-0.33	70
2 substitutions
Tangri G2	L102P,S146D	118	1	4.97	6.06	65
Tangri G3	T107D,S146D	121	1	4.67	1.02	62
BLOSUM	R110Q,S146D	119	0	4.92	0.92	65
Conservation	R76A,R110G	120	2	5.48	3.47	76
FoldX	F48D,N147D	115	2	5.07	-0.59	65
3 substitutions
Tangri G4	L102G,T107D,S146D	113	2	7.59	6.09	57
Tangri G5	L102S,T107D,S146D	115	2	7.59	4.89	58
BLOSUM	R76Q,V82T,R110Q	111	0	8.91	1.78	72
Conservation	V56E,R76A,R110G	113	3	7.95	3.29	69
FoldX	F48D,V82K,N147D	108	3	6.96	-1.18	63
4 substitutions
BLOSUM	R76Q,V82T,R110Q,S146D	104	0	10.96	2.00	65
Conservation	V56E,R76A,V82A,R110G	106	3	7.62	3.06	68
FoldX	F48D,V82E,L93E,N147D	101	4	10.01	-0.87	60
5 substitutions
BLOSUM	S71E,R76K,V82T,R110Q,S146D	98	0	12.91	1.31	62
Conservation	V56E,R76A,V82A,L91G,R110G	101	4	10.40	4.77	66
FoldX	F48D,V82E,L93E,S146D,N147D	95	4	12.05	-0.65	57
10 substitutions
BLOSUM	R53Q,S71D,R76K,V82T,S104D,	73	0	23.90	2.93	49
	R110Q,R143Q,S146D,R150Q,K154E

Variants of the entire Epo protein by Tangri et al. [22] as well as by our dynamic programming algorithm, optimizing for ProPred and allowing mutations according to BLOSUM, conservation, or FoldX. E: number of ProPred epitopes; B, C, and ΔΔG°: total substitution penalty under BLOSUM, conservation, and FoldX, resp.; S: number of SMM-align epitopes.

ISSN: 1471-2105