Improving fold resistance prediction of HIV-1 against protease and reverse transcriptase inhibitors using artificial neural networks

Abstract

BackgroundDrug resistance in HIV treatment is still a worldwide problem. Predicting resistance to antiretrovirals (ARVs) before starting any treatment is important. Prediction accuracy is essential, as low-accuracy predictions increase the risk of prescribing sub-optimal drug regimens leading to patients developing resistance sooner. Artificial Neural Networks (ANNs) are a powerful tool that would be able to assist in drug resistance prediction. In this study, we constrained the dataset to subtype B, sacrificing generalizability for a higher predictive performance, and demonstrated that the predictive quality of the ANN regression models have definite improvement for most ARVs.ResultsTrained regression ANNs were optimized for eight protease inhibitors, six nucleoside reverse transcriptase (RT) inhibitors and four non-nucleoside RT inhibitors by experimenting combinations of rare variant filtering (none versus 1 residue occurrence) and ANN topologies (1–3 hidden layers with 2, 4, 6, 8 and 10 nodes per layer). Single hidden layers (5–20 nodes) were used for training where overfitting was detected. 5-fold cross-validation produced mean R2 values over 0.95 and standard deviations lower than 0.04 for all but two antiretrovirals.ConclusionsOverall, higher accuracies and lower variances (compared to results published in 2016) were obtained by experimenting with various preprocessing methods, while focusing on the most prevalent subtype in the raw dataset (subtype B).We thus highlight the need to develop and make available subtype-specific datasets for developing higher accuracy in drug-resistance prediction methods.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-017-1782-x) contains supplementary material, which is available to authorized users.

A Computational Approach for the Prediction of HIV Resistance Based on Amino Acid and Nucleotide Descriptors

“…The prediction of HIV resistance to a drug or a combination of drugs is an important issue for the development of new potent and safe antiretroviral drugs. There are several methods aimed at predicting HIV-1 resistance and disease progression based on amino acid / nucleotide sequences of HIV core proteins (reverse transcriptase, protease) [ 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 ]. Some of them have the ultimate goal of predicting HIV-1 resistance to reverse transcriptase (RT) and protease inhibitors (PR) [ 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 ] based on amino acid or nucleotide sequences of HIV RT and PR.…”

“…Three different machine learning approaches were used to predict HIV resistance to RT and PR inhibitors [ 5 ]. Application of several different machine learning approaches to the prediction of HIV-1 resistance was also reported in References [ 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 ]. Some other approaches, including the study by Kierczak, M. et al [ 13 ], are rule-based.…”

Analysis of drug resistance in HIV protease

“…Another group reported mean R2 values of >0.95 for regression with ANN using a subset of HIVsequences restricted to subtype B with the data filtered to remove rare variants [10]. Their classification accuracy was less impressive.…”

Characterizing drug resistance using geometric ensembles from HIV protease dynamics

“…Dataset preparation: HIV subtype B protease sequence variants labeled with fold drug resistance ratios were obtained from the Stanford HIVdb unfiltered dataset 23 . These were reconstituted and filtered as explained in 24 . After ranking the sequences on the basis of decreasing average distance for each of the 8 PIs, 100 highly-resistant and 100 hyper-susceptible sequences were shortlisted, using cut-offs defined in 40 .…”