_PROBLEM CoEPrA-2006_Classification_002
 
_GROUP_NAME Curt Breneman

_GROUP_MEMBERS
Kristin Bennett
Charles Bergeron
Curt Breneman
Theresa Hepburn
Michael Krein
Min Li
Steven Mulick
Sukumar Nagamani
Matthew Sundling

_ADDRESS
Rensselaer Exploratory Centre for Cheminformatics Research,
 Rensselaer Polytechnic Institute, 110 Eighth Street,
 Troy, New York 12180-3522,
 United States of America

_MODELING_PROCEDURE
For each octapeptide, 5144 descriptors are provided. Additionally,
 147 RECON (TAE/RECON electron-density derived) descriptors were
 generated.

Each feature was centred to zero median and unit absolute deviation
 (sum of absolute difference between descriptor values and the median).

Lossless data compression was accomplished by seperately considering
 each amino acid's descriptors and the RECON descriptors. Principal
 components analysis on each descriptor subset was used to find
 components of nonzero variance. This resulted in 179 components.

Data reduction was accomplished by considering the covariance
 matrix C and an estimate S of how the covariance differs between
 the two sets. Minimum noise fractions, a generalized principal
 components analysis, was used. The generalized singular value
 decomposition of C with respect to S found "noise"-insensitive
 features. SNR is given by the singular values minus one; an SNR
 cutoff of 10 was chosen; hence, there remained 34 features.

Kernel partial least squares was used for modeling. Model selection
 found that 12 latent variables and a Gaussian kernel function
 with sigma=159.4843 was retained. A cutoff of 7.7810 was used
 to discriminate between classes. The resulting classification
 accuracy was 76.31% for leave-one-out cross-validation across the
 calibration set. Consider model f=KB+b where f is the prediction
 reported below, K is the kernel matrix, B are the regression
 coefficients and b is the bias constant. Then B and b are as follows:

B =

      -1.6374
      -9.5877
      -2.3194
       -3.141
      -21.665
     -0.24613
      -9.3587
     -0.85192
     -0.74675
     -0.61558
       7.6128
       1.8263
      -3.5451
      -2.8338
      -1.6122
      -0.4111
      -2.6264
     -0.41583
     -0.37979
       6.6286
       1.3382
       18.111
      0.52781
      -1.3566
      -9.2637
       -2.823
      -1.2057
      0.85469
       1.0383
      0.24309
      0.37763
        2.912
      -0.1232
        1.518
       1.6294
      0.67124
       2.7636
       2.3501
        -3.49
       1.2212
       2.2789
      0.32194
     -0.16696
     0.042224
    -0.086358
     -0.39148
      -1.1129
      0.62137
      0.36215
      -3.2419
      0.40171
      0.13039
       1.2837
      -3.2407
      0.92136
       1.4999
     -0.57394
       1.6376
      -1.8886
      0.20103
       1.4547
       6.1535
      -3.0457
     -0.14666
       1.5286
      0.86124
      0.86867
        -3.46
       4.2878
       3.5212
       5.6133
       2.4685
      0.99671
      0.92356
       1.0517
      0.87537


b =

       7.5509


_PREDICTION
Obj_00001 +1
Obj_00002 -1
Obj_00003 +1
Obj_00004 +1
Obj_00005 -1
Obj_00006 +1
Obj_00007 +1
Obj_00008 -1
Obj_00009 +1
Obj_00010 -1
Obj_00011 +1
Obj_00012 -1
Obj_00013 -1
Obj_00014 -1
Obj_00015 -1
Obj_00016 +1
Obj_00017 +1
Obj_00018 -1
Obj_00019 -1
Obj_00020 -1
Obj_00021 -1
Obj_00022 +1
Obj_00023 +1
Obj_00024 -1
Obj_00025 -1
Obj_00026 +1
Obj_00027 +1
Obj_00028 +1
Obj_00029 -1
Obj_00030 +1
Obj_00031 -1
Obj_00032 -1
Obj_00033 +1
Obj_00034 +1
Obj_00035 -1
Obj_00036 +1
Obj_00037 +1
Obj_00038 -1
Obj_00039 +1
Obj_00040 -1
Obj_00041 +1
Obj_00042 +1
Obj_00043 +1
Obj_00044 -1
Obj_00045 -1
Obj_00046 +1
Obj_00047 -1
Obj_00048 +1
Obj_00049 +1
Obj_00050 -1
Obj_00051 -1
Obj_00052 -1
Obj_00053 -1
Obj_00054 +1
Obj_00055 -1
Obj_00056 -1
Obj_00057 +1
Obj_00058 -1
Obj_00059 -1
Obj_00060 +1
Obj_00061 +1
Obj_00062 -1
Obj_00063 +1
Obj_00064 +1
Obj_00065 -1
Obj_00066 +1
Obj_00067 -1
Obj_00068 -1
Obj_00069 +1
Obj_00070 +1
Obj_00071 +1
Obj_00072 -1
Obj_00073 +1
Obj_00074 +1
Obj_00075 -1
Obj_00076 -1