Antigenic predicts potentially antigenic regions of a protein sequence, using the method of Kolaskar and Tongaonkar. This method is based on a single parameter and thus very simple to use.
Analysis of data from experimentally determined antigenic sites on proteins has revealed that the hydrophobic residues Cys, Leu and Val, if they occur on the surface of a protein, are more likely to be a part of antigenic sites. The method of Kolaskar and Tongaonkar to predict antigenic determinants in proteins is semi-empirical and makes use of physicochemical properties of amino acid residues and their frequencies of occurrence in experimentally known segmental epitopes.
|
Here is the default Eantigenic.dat file:
# Kolaskar AS and Tongaonkar PC (1990) FEBS Letters 276:172-174 # "A semi-emipirical method for prediction of antigenic determinants # on protein antigens" # # TABLE 1: Occurrence of amino acids in epitopes, proteins and on the surface, # and their antigenic propensity, A(p), values # # 169 antigenic determinants experimentally determined. Selected those 156 # which have less than 20 amino acids per determinant (total 2066 residues). # Calculated f(Ag) as frequency of occurrence of each residue in antigenic # determinants [f(Ag) = Epitope_occurrence/2066]. # # Used Hydrophilicity, Accessibility and Flexibility of Parker JMR, Guo D, # Hodges, RS (1986) Biochemistry 25:5425-5432. In a given protein, calculated # average for each 7-mer and assigned values to central residue of 7-mer. # Residue considered to be on the surface if any of the 7-mer values was above # the average for the protein. Used these results to get f(s) frequency of # occurrence of amino acids at the surface. # # Original table covers the 20 naturally occurring amino acids. # Values for B, Z, X use weighted averages from Edayhoff.dat # and are ignored when calculating totals # # Antigenic propensity column A(p) = f(Ag)/f(s) # # f(s) values below were back-calculated from the table in the paper # # Prediction algorithm: # # 1. calculate average propensity for each overlapping 7-mer, assign to # central residue (i+3) # # 2. calculate average for whole protein A(p)av # # 3. (a) if average for whole protein >= 1.0 then all residues having # >= 1.0 are potentially antigenic. # (b) if average for whole protein < 1.0 then all residues having # > average for whole protein (??? paper has a mangled # formula here :-) are potentially antigenic. # # 4. Find 6-mers where all residues are selected by step 3 above # # Antigenic Surface Antigenic # Amino -- Occurrence of amino acids in -- frequency frequency propensity # Acid Epitopes Surface Protein f(Ag) f(s) A(p) A 135 328 524 0.065 0.061 1.064 B 107 334 410 0.052 0.062 0.827 C 53 97 186 0.026 0.018 1.412 D 118 352 414 0.057 0.066 0.866 E 132 401 499 0.064 0.075 0.851 F 76 180 365 0.037 0.034 1.091 G 116 343 487 0.056 0.064 0.874 H 59 138 191 0.029 0.026 1.105 I 86 193 437 0.042 0.036 1.152 K 158 439 523 0.076 0.082 0.930 L 149 308 684 0.072 0.058 1.250 M 23 72 152 0.011 0.013 0.826 N 94 313 407 0.045 0.058 0.776 P 135 328 411 0.065 0.061 1.064 Q 99 252 332 0.048 0.047 1.015 R 106 314 394 0.051 0.058 0.873 S 168 429 553 0.081 0.080 1.012 T 141 401 522 0.068 0.075 0.909 V 128 239 515 0.062 0.045 1.383 W 19 55 103 0.009 0.010 0.893 X 118 306 453 0.057 0.057 1.025 Y 71 158 245 0.034 0.029 1.161 Z 119 342 433 0.058 0.064 0.916 Total 2066 5340 7944
Application of this method to a large number of proteins has shown that their method can predict antigenic determinants with about 75% accuracy which is better than most of the known methods.
Original program "ANTIGENIC" by Peter Rice (EGCG 1991)