This function performs a prediction method from 4 predicting potent compounds methods of this package. These methods are RA, EI, NN and GP.
predictChemPC(xTrain, yTrain, xTest, loghyper, method="RA")
m * n martrix of train data.
m * 1 matrix of target values consist of potencies, pIC50 or other measurements of compound affinities that are desired to be maximized.
j * n matrix of test data.
3 * 1 matrix of loghyper parameters which is the output of trainChemPC function.
One of "EI", "GP", "NN" or "RA".
This function withholds 4 methods to predict potent compounds.
method is one of:
EI A compound for which maximum expected potency improvement is reached.
GP A compound holding maximum predicted potency in test data is selected.
NN A compound that is nearest (Tonimito Coefficient as distance measure) to the most potent compound in training data is selected.
RA As it's name suggests, a compound is selected randomly.
Feature selection employed in this package is based on Spearman Rank Correlation such that
before each training step those attributes in which revealed a significant Spearman rank correlation
with the logarithmic potency values (q-value < 5
are computed from original p-values via the multiple testing correction method by Benjamini and Hochberg.
It returns index of most potent compound in original test set w.r.t. selected method.
1.Predicting Potent Compounds via Model-Based Global Optimization, Journal of Chemical Information and Modeling, 2013, 53 (3), pp 553-559, M Ahmadi, M Vogt, P Iyer, J Bajorath, H Froehlich.
2.Software MOE is used to calculate the numerical descriptors in data sets. Ref: http://www.chemcomp.com/MOE-Molecular_Operating_Environment.htm
3.ChEMBL was the source of the compound data and potency annotations in data sets. Ref: https://www.ebi.ac.uk/chembl/
1 2 3 4 5