A comparison of three computational modelling methods for the prediction of virological response to combination HIV therapy

Artificial Intelligence and Machine Learning Based Prediction of Viral Load and CD4 Status of People Living with HIV (PLWH) on Anti-Retroviral Treatment in Gedeo Zone Public Hospitals

Background

Despite the success made in scaling up HIV treatment activities, there remains a tremendous unmet demand for the monitoring of the disease progression and treatment success, which threatens HIV/AIDS treatment and control. This research presented the assessments of viral load and CD4 classification of adults enrolled in ART care using machine learning algorithms.

Methods

We trained, validated, and tested eight machine learning (ML) classifier algorithms with historical data, including demographics, clinical, and laboratory data. Data were extracted from the ART registry database of Yirgacheffe Primary Hospital and Dilla University Referral Hospital. ML classifiers were trained to predict virological failure (viral load >1000 copies/mL) and poor CD4 (CD4 cell count <200 cells/mL). The model predictive performances were evaluated using accuracy, sensitivity, specificity, precision, f1-score, F-beta scores, and AUC.

Results

The mean age of the sample participants was 41.6 years (SD = 10.9). The experimental results showed that XGB classifier ranked as the best algorithm for viral load prediction in terms of sensitivity (97%), f1-score (96%), AUC (0.99), accuracy (96%), followed by RF. The GB classifier exhibited a better predictive capability in predicting participants with a CD4 cell count <200 cells/mL.

Conclusion

In this study, the XGB and RF models had the highest accuracy and outperformed on various evaluation metrics among the models examined for viral load classification. In the prediction of participants CD4, GB model had the highest accuracy.

Prediction of HIV-1 protease resistance using genotypic, phenotypic, and molecular information with artificial neural networks

Drug resistance is a primary barrier to effective treatments of HIV/AIDS. Calculating quantitative relations between genotype and phenotype observations for each inhibitor with cell-based assays requires time and money-consuming experiments. Machine learning models are good options for tackling these problems by generalizing the available data with suitable linear or nonlinear mappings. The main aim of this study is to construct drug isolate fold (DIF) change-based artificial neural network (ANN) models for estimating the resistance potential of molecules inhibiting the HIV-1 protease (PR) enzyme. Throughout the study, seven of eight protease inhibitors (PIs) have been included in the training set and the remaining ones in the test set. We have obtained 11,803 genotype-phenotype data points for eight PIs from Stanford HIV drug resistance database. Using the leave-one-out (LVO) procedure, eight ANN models have been produced to measure the learning capacity of models from the descriptors of the inhibitors. Mean R² value of eight ANN models for unseen inhibitors is 0.716, and the 95% confidence interval (CI) is [0.592-0.840]. Predicting the fold change resistance for hundreds of isolates allowed a robust comparison of drug pairs. These eight models have predicted the drug resistance tendencies of each inhibitor pair with the mean 2D correlation coefficient of 0.933 and 95% CI [0.930-0.938]. A classification problem has been created to predict the ordered relationship of the PIs, and the mean accuracy, sensitivity, specificity, and Matthews correlation coefficient (MCC) values are calculated as 0.954, 0.791, 0.791, and 0.688, respectively. Furthermore, we have created an external test dataset consisting of 51 unique known HIV-1 PR inhibitors and 87 genotype-phenotype relations. Our developed ANN model has accuracy and area under the curve (AUC) values of 0.749 and 0.818 to predict the ordered relationships of molecules on the same strain for the external dataset. The currently derived ANN models can accurately predict the drug resistance tendencies of PI pairs. This observation could help test new inhibitors with various isolates.

A Study on a Neural Network Risk Simulation Model Construction for Avian Influenza A (H7N9) Outbreaks in Humans in China during 2013-2017

The main purposes of this study were to explore the spatial distribution characteristics of H7N9 human infections during 2013-2017, and to construct a neural network risk simulation model of H7N9 outbreaks in China and evaluate their effects. First, ArcGIS 10.6 was used for spatial autocorrelation analysis, and cluster patterns ofH7N9 outbreaks were analyzed in China during 2013-2017 to detect outbreaks' hotspots. During the study period, the incidence of H7N9 outbreaks in China was high in the eastern and southeastern coastal areas of China, with a tendency to spread to the central region. Moran's I values of global spatial autocorrelation of H7N9 outbreaks in China from 2013 to 2017 were 0.080128, 0.073792, 0.138015, 0.139221 and 0.050739, respectively (p < 0.05) indicating a statistically significant positive correlation of the epidemic. Then, SPSS 20.0 was used to analyze the correlation between H7N9 outbreaks in China and population, livestock production, the distance between the case and rivers, poultry farming, poultry market, vegetation index, etc. Statistically significant influencing factors screened out by correlation analysis were population of the city, average vegetation of the city, and the distance between the case and rivers (p < 0.05), which were included in the neural network risk simulation model of H7N9 outbreaks in China. The simulation accuracy of the neural network risk simulation model of H7N9 outbreaks in China from 2013 to 2017 were 85.71%, 91.25%, 91.54%, 90.49% and 92.74%, and the AUC were 0.903, 0.976, 0.967, 0.963 and 0.970, respectively, showing a good simulation effect of H7N9 epidemics in China. The innovation of this study lies in the epidemiological study of H7N9 outbreaks by using a variety of technical means, and the construction of a neural network risk simulation model of H7N9 outbreaks in China. This study could provide valuable references for the prevention and control of H7N9 outbreaks in China.

Studying patterns and predictors of HIV viral suppression using A Big Data approach: a research protocol

Background

Given the importance of viral suppression in ending the HIV epidemic in the US and elsewhere, an optimal predictive model of viral status can help clinicians identify those at risk of poor viral control and inform clinical improvements in HIV treatment and care. With an increasing availability of electronic health record (EHR) data and social environmental information, there is a unique opportunity to improve our understanding of the dynamic pattern of viral suppression. Using a statewide cohort of people living with HIV (PLWH) in South Carolina (SC), the overall goal of the proposed research is to examine the dynamic patterns of viral suppression, develop optimal predictive models of various viral suppression indicators, and translate the models to a beta version of service-ready tools for clinical decision support.

Methods

The PLWH cohort will be identified through the SC Enhanced HIV/AIDS Reporting System (eHARS). The SC Office of Revenue and Fiscal Affairs (RFA) will extract longitudinal EHR clinical data of all PLWH in SC from multiple health systems, obtain data from other state agencies, and link the patient-level data with county-level data from multiple publicly available data sources. Using the deidentified data, the proposed study will consist of three operational phases: Phase 1: “Pattern Analysis” to identify the longitudinal dynamics of viral suppression using multiple viral load indicators; Phase 2: “Model Development” to determine the critical predictors of multiple viral load indicators through artificial intelligence (AI)-based modeling accounting for multilevel factors; and Phase 3: “Translational Research” to develop a multifactorial clinical decision system based on a risk prediction model to assist with the identification of the risk of viral failure or viral rebound when patients present at clinical visits.

Discussion

With both extensive data integration and data analytics, the proposed research will: (1) improve the understanding of the complex inter-related effects of longitudinal trajectories of HIV viral suppressions and HIV treatment history while taking into consideration multilevel factors; and (2) develop empirical public health approaches to achieve ending the HIV epidemic through translating the risk prediction model to a multifactorial decision system that enables the feasibility of AI-assisted clinical decisions.

Development and internal validation of the HIV In-hospital Mortality Prediction (HIV-IMP) risk score

BACKGROUND

Despite advances in availability and access to antiretroviral therapy (ART), HIV still ranks as a major cause of global mortality. Hence, the aim of this study was to develop and internally validate a risk score capable of accurately predicting in-hospital mortality in HIV-positive patients requiring hospital admission.

METHODS

Consecutive HIV-positive patients presenting to the Charlotte Maxeke Johannesburg Academic Hospital adult emergency department between 7 July 2017 and 18 October 2018 were prospectively enrolled. Multivariate logistic regression was used to determine parameters for inclusion in the final risk score. Discrimination and calibration were assessed by means of the area under the receiver operating curve (AUROC) and the Hosmer-Lemeshow goodness-of-fit test, respectively. Internal validation was conducted using the regular bootstrap technique.

RESULTS

The overall in-hospital mortality rate was 13.6% (n = 166). Eight predictors were included in the final risk score: ART non-adherence or not yet on ART, Glasgow Coma Scale < 15, respiratory rate > 20 breaths/min, oxygen saturation < 90%, white cell count < 4 × 10⁹ /L, creatinine > 120 μmol/L, lactate > 2 mmol/L and albumin < 35 g/L. After internal validation, the risk score maintained good discrimination [AUROC 0.83, 95% confidence interval (CI): 0.78-0.88] and calibration (Hosmer-Lemeshow χ² = 2.26, p = 0.895).

CONCLUSION

The HIV In-hospital Mortality Prediction (HIV-IMP) risk score has overall good discrimination and calibration and is relatively easy to use. Further studies should be aimed at externally validating the score in varying clinical settings.

Artificial intelligence and machine learning assisted drug delivery for effective treatment of infectious diseases

In the era of antimicrobial resistance, the prevalence of multidrug-resistant microorganisms that resist conventional antibiotic treatment has steadily increased. Thus, it is now unquestionable that infectious diseases are significant global burdens that urgently require innovative treatment strategies. Emerging studies have demonstrated that artificial intelligence (AI) can transform drug delivery to promote effective treatment of infectious diseases. In this review, we propose to evaluate the significance, essential principles, and popular tools of AI in drug delivery for infectious disease treatment. Specifically, we will focus on the achievements and key findings of current research, as well as the applications of AI on drug delivery throughout the whole antimicrobial treatment process, with an emphasis on drug development, treatment regimen optimization, drug delivery system and administration route design, and drug delivery outcome prediction. To that end, the challenges of AI in drug delivery for infectious disease treatments and their current solutions and future perspective will be presented and discussed.

2021 update to HIV-TRePS: a highly flexible and accurate system for the prediction of treatment response from incomplete baseline information in different healthcare settings

OBJECTIVES

With the goal of facilitating the use of HIV-TRePS to optimize therapy in settings with limited healthcare resources, we aimed to develop computational models to predict treatment responses accurately in the absence of commonly used baseline data.

METHODS

Twelve sets of random forest models were trained using very large, global datasets to predict either the probability of virological response (classifier models) or the absolute change in viral load in response to a new regimen (absolute models) following virological failure. Two 'standard' models were developed with all baseline variables present and 10 others developed without HIV genotype, time on therapy, CD4 count or any combination of the above.

RESULTS

The standard classifier models achieved an AUC of 0.89 in cross-validation and independent testing. Models with missing variables achieved AUC values of 0.78-0.90. The standard absolute models made predictions that correlated significantly with observed changes in viral load with a mean absolute error of 0.65 log10 copies HIV RNA/mL in cross-validation and 0.69 log10 copies HIV RNA/mL in independent testing. Models with missing variables achieved values of 0.65-0.75 log10 copies HIV RNA/mL. All models identified alternative regimens that were predicted to be effective for the vast majority of cases where the new regimen prescribed in the clinic failed. All models were significantly better predictors of treatment response than genotyping with rules-based interpretation.

CONCLUSIONS

These latest models that predict treatment responses accurately, even when a number of baseline variables are not available, are a major advance with greatly enhanced potential benefit, particularly in resource-limited settings. The only obstacle to realizing this potential is the willingness of healthcare professions to use the system.

Predicting Virological Response to HIV Treatment Over Time: A Tool for Settings With Different Definitions of Virological Response

OBJECTIVE

Definitions of virological response vary from <50 up to 1000 copies of HIV-RNA/mL. Our previous models estimate the probability of HIV drug combinations reducing the viral load to <50 copies/mL, with no indication of whether higher thresholds of response may be achieved. Here, we describe the development of models that predict absolute viral load over time.

METHODS

Two sets of random forest models were developed using 50,270 treatment change episodes from more than 20 countries. The models estimated viral load at different time points following the introduction of a new regimen from variables including baseline viral load, CD4 count, and treatment history. One set also used genotypes in their predictions. Independent data sets were used for evaluation.

RESULTS

Both models achieved highly significant correlations between predicted and actual viral load changes (r = 0.67-0.68, mean absolute error of 0.73-0.74 log10 copies/mL). The models produced curves of virological response over time. Using failure definitions of <100, 400, or 1000 copies/mL, but not 50 copies/mL, both models were able to identify alternative regimens they predicted to be effective for the majority of cases where the new regimen prescribed in the clinic failed.

CONCLUSIONS

These models could be useful for selecting the optimum combination therapy for patients requiring a change in therapy in settings using any definition of virological response. They also give an idea of the likely response curve over time. Given that genotypes are not required, these models could be a useful addition to the HIV-TRePS system for those in resource-limited settings.

Using machine learning methods to determine a typology of patients with HIV-HCV infection to be treated with antivirals

Several European countries have established criteria for prioritising initiation of treatment in patients infected with the hepatitis C virus (HCV) by grouping patients according to clinical characteristics. Based on neural network techniques, our objective was to identify those factors for HIV/HCV co-infected patients (to which clinicians have given careful consideration before treatment uptake) that have not being included among the prioritisation criteria. This study was based on the Spanish HERACLES cohort (NCT02511496) (April-September 2015, 2940 patients) and involved application of different neural network models with different basis functions (product-unit, sigmoid unit and radial basis function neural networks) for automatic classification of patients for treatment. An evolutionary algorithm was used to determine the architecture and estimate the coefficients of the model. This machine learning methodology found that radial basis neural networks provided a very simple model in terms of the number of patient characteristics to be considered by the classifier (in this case, six), returning a good overall classification accuracy of 0.767 and a minimum sensitivity (for the classification of the minority class, untreated patients) of 0.550. Finally, the area under the ROC curve was 0.802, which proved to be exceptional. The parsimony of the model makes it especially attractive, using just eight connections. The independent variable “recent PWID” is compulsory due to its importance. The simplicity of the model means that it is possible to analyse the relationship between patient characteristics and the probability of belonging to the treated group.

Artificial Intelligence in Drug Treatment

The most common applications of artificial intelligence (AI) in drug treatment have to do with matching patients to their optimal drug or combination of drugs, predicting drug-target or drug-drug interactions, and optimizing treatment protocols. This review outlines some of the recently developed AI methods aiding the drug treatment and administration process. Selection of the best drug(s) for a patient typically requires the integration of patient data, such as genetics or proteomics, with drug data, like compound chemical descriptors, to score the therapeutic efficacy of drugs. The prediction of drug interactions often relies on similarity metrics, assuming that drugs with similar structures or targets will have comparable behavior or may interfere with each other. Optimizing the dosage schedule for administration of drugs is performed using mathematical models to interpret pharmacokinetic and pharmacodynamic data. The recently developed and powerful models for each of these tasks are addressed, explained, and analyzed here.

A comparative study of logistic regression based machine learning techniques for prediction of early virological suppression in antiretroviral initiating HIV patients

Background

Treatment with effective antiretroviral therapy (ART) lowers morbidity and mortality among HIV positive individuals. Effective highly active antiretroviral therapy (HAART) should lead to undetectable viral load within 6 months of initiation of therapy. Failure to achieve and maintain viral suppression may lead to development of resistance and increase the risk of viral transmission. In this paper three logistic regression based machine learning approaches are developed to predict early virological outcomes using easily measurable baseline demographic and clinical variables (age, body weight, sex, TB disease status, ART regimen, viral load, CD4 count). The predictive performance and generalizability of the approaches are compared.

Methods

The multitask temporal logistic regression (MTLR), patient specific survival prediction (PSSP) and simple logistic regression (SLR) models were developed and validated using the IDI research cohort data and predictive performance tested on an external dataset from the EFV cohort. The model calibration and discrimination plots, discriminatory measures (AUROC, F1) and overall predictive performance (brier score) were assessed.

Results

The MTLR model outperformed the PSSP and SLR models in terms of goodness of fit (RMSE = 0.053, 0.1, and 0.14 respectively), discrimination (AUROC = 0.92, 0.75 and 0.53 respectively) and general predictive performance (Brier score= 0.08, 0.19, 0.11 respectively). The predictive importance of variables varied with time after initiation of ART. The final MTLR model accurately (accuracy = 92.9%) predicted outcomes in the external (EFV cohort) dataset with satisfactory discrimination (0.878) and a low (6.9%) false positive rate.

Conclusion

Multitask Logistic regression based models are capable of accurately predicting early virological suppression using readily available baseline demographic and clinical variables and could be used to derive a risk score for use in resource limited settings.

2018 update to the HIV-TRePS system: the development of new computational models to predict HIV treatment outcomes, with or without a genotype, with enhanced usability for low-income settings

Objectives

Optimizing antiretroviral drug combination on an individual basis can be challenging, particularly in settings with limited access to drugs and genotypic resistance testing. Here we describe our latest computational models to predict treatment responses, with or without a genotype, and compare their predictive accuracy with that of genotyping.

Methods

Random forest models were trained to predict the probability of virological response to a new therapy introduced following virological failure using up to 50 000 treatment change episodes (TCEs) without a genotype and 18 000 TCEs including genotypes. Independent data sets were used to evaluate the models. This study tested the effects on model accuracy of relaxing the baseline data timing windows, the use of a new filter to exclude probable non-adherent cases and the addition of maraviroc, tipranavir and elvitegravir to the system.

Results

The no-genotype models achieved area under the receiver operator characteristic curve (AUC) values of 0.82 and 0.81 using the standard and relaxed baseline data windows, respectively. The genotype models achieved AUC values of 0.86 with the new non-adherence filter and 0.84 without. Both sets of models were significantly more accurate than genotyping with rules-based interpretation, which achieved AUC values of only 0.55-0.63, and were marginally more accurate than previous models. The models were able to identify alternative regimens that were predicted to be effective for the vast majority of cases in which the new regimen prescribed in the clinic failed.

Conclusions

These latest global models predict treatment responses accurately even without a genotype and have the potential to help optimize therapy, particularly in resource-limited settings.

Prediction of HIV Drug Resistance by Combining Sequence and Structural Properties

Drug resistance is a major obstacle faced by therapist in treating HIV infected patients. The reason behind these phenomena is either protein mutation or the changes in gene expression level that induces resistance to drug treatments. These mutations affect the drug binding activity, hence resulting in failure of treatment. Therefore, it is necessary to conduct resistance testing in order to carry out HIV effective therapy. This study combines both sequence and structural features for predicting HIV resistance by applying SVM and Random Forests classifiers. The model was tested on the mutants of HIV-1 protease and reverse transcriptase. Taken together the features we have used in our method, total contact energies among multiple mutations have a strong impact in predicting resistance as they are crucial in understanding the interactions of HIV mutants. The combination of sequence-structure features offers high accuracy with support vector machines as compared to Random Forests classifier. Both single and acquisition of multiple mutations are important in predicting HIV resistance to certain drug treatments. We have discovered the practicality of these features; hence, these can be used in the future to predict resistance for other complex diseases.

Predicting CD4 count changes among patients on antiretroviral treatment: Application of data mining techniques

BACKGROUND AND OBJECTIVES

To monitor the progress of therapy and disease progression, periodic CD4 counts are required throughout the course of HIV/AIDS care and support. The demand for CD4 count measurement is increasing as ART programs expand over the last decade. This study aimed to predict CD4 count changes and to identify the predictors of CD4 count changes among patients on ART.

METHODS

A cross-sectional study was conducted at the University of Gondar Hospital from 3,104 adult patients on ART with CD4 counts measured at least twice (baseline and most recent). Data were retrieved from the HIV care clinic electronic database and patients` charts. Descriptive data were analyzed by SPSS version 20. Cross-Industry Standard Process for Data Mining (CRISP-DM) methodology was followed to undertake the study. WEKA version 3.8 was used to conduct a predictive data mining. Before building the predictive data mining models, information gain values and correlation-based Feature Selection methods were used for attribute selection. Variables were ranked according to their relevance based on their information gain values. J48, Neural Network, and Random Forest algorithms were experimented to assess model accuracies.

RESULT

The median duration of ART was 191.5 weeks. The mean CD4 count change was 243 (SD 191.14) cells per microliter. Overall, 2427 (78.2%) patients had their CD4 counts increased by at least 100 cells per microliter, while 4% had a decline from the baseline CD4 value. Baseline variables including age, educational status, CD8 count, ART regimen, and hemoglobin levels predicted CD4 count changes with predictive accuracies of J48, Neural Network, and Random Forest being 87.1%, 83.5%, and 99.8%, respectively. Random Forest algorithm had a superior performance accuracy level than both J48 and Artificial Neural Network. The precision, sensitivity and recall values of Random Forest were also more than 99%.

CONCLUSIONS

Nearly accurate prediction results were obtained using Random Forest algorithm. This algorithm could be used in a low-resource setting to build a web-based prediction model for CD4 count changes.

A survey of machine learning applications in HIV clinical research and care

A wealth of genetic, demographic, clinical and biomarker data is collected from routine clinical care of HIV patients and exists in the form of medical records available among the medical care and research communities. Machine learning (ML) methods have the ability to identify and discover patterns in complex datasets and predict future outcomes of HIV treatment. We survey published studies that make use of ML techniques in HIV clinical research and care. An advanced search relevant to the use of ML in HIV research was conducted in the PubMed biomedical database. The survey outcomes of interest include data sources, ML techniques, ML tasks and ML application paradigms. A growing trend in application of ML in HIV research was observed. The application paradigm has diversified to include practical clinical application, but statistical analysis remains the most dominant application. There is an increase in the use of genomic sources of data and high performance non-parametric ML methods with a focus on combating resistance to antiretroviral therapy (ART). There is need for improvement in collection of health records data and increased training in ML so as to translate ML research into clinical application in HIV management.

An update to the HIV-TRePS system: the development and evaluation of new global and local computational models to predict HIV treatment outcomes, with or without a genotype

OBJECTIVES

Optimizing antiretroviral drug combination on an individual basis in resource-limited settings is challenging because of the limited availability of drugs and genotypic resistance testing. Here, we describe our latest computational models to predict treatment responses, with or without a genotype, and compare the potential utility of global and local models as a treatment tool for South Africa.

METHODS

Global random forest models were trained to predict the probability of virological response to therapy following virological failure using 29 574 treatment change episodes (TCEs) without a genotype, 3179 of which were from South Africa and were used to develop local models. In addition, 15 130 TCEs including genotypes were used to develop another set of models. The 'no-genotype' models were tested with an independent global test set (n = 1700) plus a subset from South Africa (n = 222). The genotype models were tested with 750 independent cases.

RESULTS

The global no-genotype models achieved area under the receiver-operating characteristic curve (AUC) values of 0.82 and 0.79 with the global and South African tests sets, respectively, and the South African models achieved AUCs of 0.70 and 0.79. The genotype models achieved an AUC of 0.84. The global no-genotype models identified more alternative, locally available regimens that were predicted to be effective for cases that failed their new regimen in the South African clinics than the local models. Both sets of models were significantly more accurate predictors of outcomes than genotyping with rules-based interpretation.

CONCLUSIONS

These latest global models predict treatment responses accurately even without a genotype, out-performed the local South African models and have the potential to help optimize therapy, particularly in resource-limited settings.

Computational models as predictors of HIV treatment outcomes for the Phidisa cohort in South Africa

BACKGROUND

Selecting the optimal combination of HIV drugs for an individual in resource-limited settings is challenging because of the limited availability of drugs and genotyping.

OBJECTIVE

The evaluation as a potential treatment support tool of computational models that predict response to therapy without a genotype, using cases from the Phidisa cohort in South Africa.

METHODS

Cases from Phidisa of treatment change following failure were identified that had the following data available: baseline CD4 count and viral load, details of failing and previous antiretroviral drugs, drugs in new regimen and time to follow-up. The HIV Resistance Response Database Initiative's (RDI's) models used these data to predict the probability of a viral load < 50 copies/mL at follow-up. The models were also used to identify effective alternative combinations of three locally available drugs.

RESULTS

The models achieved accuracy (area under the receiver-operator characteristic curve) of 0.72 when predicting response to therapy, which is less accurate than for an independent global test set (0.80) but at least comparable to that of genotyping with rules-based interpretation. The models were able to identify alternative locally available three-drug regimens that were predicted to be effective in 69% of all cases and 62% of those whose new treatment failed in the clinic.

CONCLUSION

The predictive accuracy of the models for these South African patients together with the results of previous studies suggest that the RDI's models have the potential to optimise treatment selection and reduce virological failure in different patient populations, without the use of a genotype.

Prediction of HIV drug resistance from genotype with encoded three-dimensional protein structure

BACKGROUND

Drug resistance has become a severe challenge for treatment of HIV infections. Mutations accumulate in the HIV genome and make certain drugs ineffective. Prediction of resistance from genotype data is a valuable guide in choice of drugs for effective therapy.

RESULTS

In order to improve the computational prediction of resistance from genotype data we have developed a unified encoding of the protein sequence and three-dimensional protein structure of the drug target for classification and regression analysis. The method was tested on genotype-resistance data for mutants of HIV protease and reverse transcriptase. Our graph based sequence-structure approach gives high accuracy with a new sparse dictionary classification method, as well as support vector machine and artificial neural networks classifiers. Cross-validated regression analysis with the sparse dictionary gave excellent correlation between predicted and observed resistance.

CONCLUSION

The approach of encoding the protein structure and sequence as a 210-dimensional vector, based on Delaunay triangulation, has promise as an accurate method for predicting resistance from sequence for drugs inhibiting HIV protease and reverse transcriptase.

An update to the HIV-TRePS system: the development of new computational models that do not require a genotype to predict HIV treatment outcomes

OBJECTIVES

The optimal individualized selection of antiretroviral drugs in resource-limited settings is challenging because of the limited availability of drugs and genotyping. Here we describe the development of the latest computational models to predict the response to combination antiretroviral therapy without a genotype, for potential use in such settings.

METHODS

Random forest models were trained to predict the probability of a virological response to therapy (<50 copies HIV RNA/mL) following virological failure using the following data from 22,567 treatment-change episodes including 1090 from southern Africa: baseline viral load and CD4 cell count, treatment history, drugs in the new regimen, time to follow-up and follow-up viral load. The models were assessed during cross-validation and with an independent global test set of 1000 cases including 100 from southern Africa. The models' accuracy [area under the receiver-operating characteristic curve (AUC)] was evaluated and compared with genotyping using rules-based interpretation systems for those cases with genotypes available.

RESULTS

The models achieved AUCs of 0.79-0.84 (mean 0.82) during cross-validation, 0.80 with the global test set and 0.78 with the southern African subset. The AUCs were significantly lower (0.56-0.57) for genotyping.

CONCLUSIONS

The models predicted virological response to HIV therapy without a genotype as accurately as previous models that included a genotype. They were accurate for cases from southern Africa and significantly more accurate than genotyping. These models will be accessible via the online treatment support tool HIV-TRePS and have the potential to help optimize antiretroviral therapy in resource-limited settings where genotyping is not generally available.

Computational models can predict response to HIV therapy without a genotype and may reduce treatment failure in different resource-limited settings

OBJECTIVES

Genotypic HIV drug-resistance testing is typically 60%-65% predictive of response to combination antiretroviral therapy (ART) and is valuable for guiding treatment changes. Genotyping is unavailable in many resource-limited settings (RLSs). We aimed to develop models that can predict response to ART without a genotype and evaluated their potential as a treatment support tool in RLSs.

METHODS

Random forest models were trained to predict the probability of response to ART (≤400 copies HIV RNA/mL) using the following data from 14 891 treatment change episodes (TCEs) after virological failure, from well-resourced countries: viral load and CD4 count prior to treatment change, treatment history, drugs in the new regimen, time to follow-up and follow-up viral load. Models were assessed by cross-validation during development, with an independent set of 800 cases from well-resourced countries, plus 231 cases from Southern Africa, 206 from India and 375 from Romania. The area under the receiver operating characteristic curve (AUC) was the main outcome measure.

RESULTS

The models achieved an AUC of 0.74-0.81 during cross-validation and 0.76-0.77 with the 800 test TCEs. They achieved AUCs of 0.58-0.65 (Southern Africa), 0.63 (India) and 0.70 (Romania). Models were more accurate for data from the well-resourced countries than for cases from Southern Africa and India (P < 0.001), but not Romania. The models identified alternative, available drug regimens predicted to result in virological response for 94% of virological failures in Southern Africa, 99% of those in India and 93% of those in Romania.

CONCLUSIONS

We developed computational models that predict virological response to ART without a genotype with comparable accuracy to genotyping with rule-based interpretation. These models have the potential to help optimize antiretroviral therapy for patients in RLSs where genotyping is not generally available.

What do molecules do when we are not looking? State sequence analysis for stochastic chemical systems

Many biomolecular systems depend on orderly sequences of chemical transformations or reactions. Yet, the dynamics of single molecules or small-copy-number molecular systems are significantly stochastic. Here, we propose state sequence analysis--a new approach for predicting or visualizing the behaviour of stochastic molecular systems by computing maximum probability state sequences, based on initial conditions or boundary conditions. We demonstrate this approach by analysing the acquisition of drug-resistance mutations in the human immunodeficiency virus genome, which depends on rare events occurring on the time scale of years, and the stochastic opening and closing behaviour of a single sodium ion channel, which occurs on the time scale of milliseconds. In both cases, we find that our approach yields novel insights into the stochastic dynamical behaviour of these systems, including insights that are not correctly reproduced in standard time-discretization approaches to trajectory analysis.

Structure-based methods for predicting target mutation-induced drug resistance and rational drug design to overcome the problem

Drug resistance has become one of the biggest challenges in drug discovery and/or development and has attracted great research interests worldwide. During the past decade, computational strategies have been developed to predict target mutation-induced drug resistance. Meanwhile, various molecular design strategies, including targeting protein backbone, targeting highly conserved residues and dual/multiple targeting, have been used to design novel inhibitors for combating the drug resistance. In this article we review recent advances in development of computational methods for target mutation-induced drug resistance prediction and strategies for rational design of novel inhibitors that could be effective against the possible drug-resistant mutants of the target.

The use of computational models to predict response to HIV therapy for clinical cases in Romania

INTRODUCTION

A major challenge in Romania is the optimisation of antiretroviral therapy for the many HIV-infected adults with, on average, a decade of treatment experience. The RDI has developed computational models that predict virological response to therapy but these require a genotype, which is not routinely available in Romania. Moreover the models, which were trained without any Romanian data, have proved most accurate for patients from the healthcare settings that contributed the training data. Here we develop and test a novel model that does not require a genotype, with test data from Romania.

METHODS

A random forest (RF) model was developed to predict the probability of the HIV viral load (VL) being reduced to <50 copies/ml following therapy change. The input variables were baseline VL, CD4 count, treatment history and time to follow-up. The model was developed with 3188 treatment changes episodes (TCEs) from North America, Western Europe and Australia. The model's predictions for 100 independent TCEs from the RDI database were compared to those of a model trained with the same data plus genotypes and then tested using 39 TCEs from Romania in terms of the area under the ROC curve (AUC).

RESULTS

When tested with the 100 independent RDI TCEs, the AUC values for the models with and without genotypes were 0.88 and 0.86 respectively. For the 39 Romanian TCEs the AUC was 0.60. However, when 14 cases with viral loads that may have been between 50 and 400 copies were removed, the AUC increased to 0.83.

DISCUSSION

Despite having been trained without data from Romania, the model predicted treatment responses in treatment-experienced Romanian patients with clade F virus accurately without the need for a genotype. The results suggest that this approach might be generalisable and useful in helping design optimal salvage regimens for treatment-experienced patients in countries with limited resources where genotyping is not always available.

Predictive models for maximum recommended therapeutic dose of antiretroviral drugs

A novel method for predicting maximum recommended therapeutic dose (MRTD) is presented using quantitative structure property relationships (QSPRs) and artificial neural networks (ANNs). MRTD data of 31 structurally diverse Antiretroviral drugs (ARVs) were collected from FDA MRTD Database or package inserts. Molecular property descriptors of each compound, that is, molecular mass, aqueous solubility, lipophilicity, biotransformation half life, oxidation half life, and biodegradation probability were calculated from their SMILES codes. A training set (n = 23) was used to construct multiple linear regression and back propagation neural network models. The models were validated using an external test set (n = 8) which demonstrated that MRTD values may be predicted with reasonable accuracy. Model predictability was described by root mean squared errors (RMSEs), Kendall's correlation coefficients (tau), P-values, and Bland Altman plots for method comparisons. MRTD was predicted by a 6-3-1 neural network model (RMSE = 13.67, tau = 0.643, P = 0.035) more accurately than by the multiple linear regression (RMSE = 27.27, tau = 0.714, P = 0.019) model. Both models illustrated a moderate correlation between aqueous solubility of antiretroviral drugs and maximum therapeutic dose. MRTD prediction may assist in the design of safer, more effective treatments for HIV infection.

A multifaceted analysis of HIV-1 protease multidrug resistance phenotypes

Background

Great strides have been made in the effective treatment of HIV-1 with the development of second-generation protease inhibitors (PIs) that are effective against historically multi-PI-resistant HIV-1 variants. Nevertheless, mutation patterns that confer decreasing susceptibility to available PIs continue to arise within the population. Understanding the phenotypic and genotypic patterns responsible for multi-PI resistance is necessary for developing PIs that are active against clinically-relevant PI-resistant HIV-1 variants.

Results

In this work, we use globally optimal integer programming-based clustering techniques to elucidate multi-PI phenotypic resistance patterns using a data set of 398 HIV-1 protease sequences that have each been phenotyped for susceptibility toward the nine clinically-approved HIV-1 PIs. We validate the information content of the clusters by evaluating their ability to predict the level of decreased susceptibility to each of the available PIs using a cross validation procedure. We demonstrate the finding that as a result of phenotypic cross resistance, the considered clinical HIV-1 protease isolates are confined to ~6% or less of the clinically-relevant phenotypic space. Clustering and feature selection methods are used to find representative sequences and mutations for major resistance phenotypes to elucidate their genotypic signatures. We show that phenotypic similarity does not imply genotypic similarity, that different PI-resistance mutation patterns can give rise to HIV-1 isolates with similar phenotypic profiles.

Conclusion

Rather than characterizing HIV-1 susceptibility toward each PI individually, our study offers a unique perspective on the phenomenon of PI class resistance by uncovering major multidrug-resistant phenotypic patterns and their often diverse genotypic determinants, providing a methodology that can be applied to understand clinically-relevant phenotypic patterns to aid in the design of novel inhibitors that target other rapidly evolving molecular targets as well.

The development of an expert system to predict virological response to HIV therapy as part of an online treatment support tool

OBJECTIVE

The optimum selection and sequencing of combination antiretroviral therapy to maintain viral suppression can be challenging. The HIV Resistance Response Database Initiative has pioneered the development of computational models that predict the virological response to drug combinations. Here we describe the development and testing of random forest models to power an online treatment selection tool.

METHODS

Five thousand, seven hundred and fifty-two treatment change episodes were selected to train a committee of 10 models to predict the probability of virological response to a new regimen. The input variables were antiretroviral treatment history, baseline CD4 cell count, viral load and genotype, drugs in the new regimen, time from treatment change to follow-up and follow-up viral load values. The models were assessed during cross-validation and with an independent set of 50 treatment change episodes by plotting receiver-operator characteristic curves and their performance compared with genotypic sensitivity scores from rules-based genotype interpretation systems.

RESULTS

The models achieved an area under the curve during cross-validation of 0.77-0.87 (mean = 0.82), accuracy of 72-81% (mean = 77%), sensitivity of 62-80% (mean = 67%) and specificity of 75-89% (mean = 81%). When tested with the 50 test cases, the area under the curve was 0.70-0.88, accuracy 64-82%, sensitivity 62-80% and specificity 68-95%. The genotypic sensitivity scores achieved an area under the curve of 0.51-0.52, overall accuracy of 54-56%, sensitivity of 43-64% and specificity of 41-73%.

CONCLUSION

The models achieved a consistent, high level of accuracy in predicting treatment responses, which was markedly superior to that of genotypic sensitivity scores. The models are being used to power an experimental system now available via the Internet.

A prognostic model for estimating the time to virologic failure in HIV-1 infected patients undergoing a new combination antiretroviral therapy regimen

Background

HIV-1 genotypic susceptibility scores (GSSs) were proven to be significant prognostic factors of fixed time-point virologic outcomes after combination antiretroviral therapy (cART) switch/initiation. However, their relative-hazard for the time to virologic failure has not been thoroughly investigated, and an expert system that is able to predict how long a new cART regimen will remain effective has never been designed.

Methods

We analyzed patients of the Italian ARCA cohort starting a new cART from 1999 onwards either after virologic failure or as treatment-naïve. The time to virologic failure was the endpoint, from the 90^th day after treatment start, defined as the first HIV-1 RNA > 400 copies/ml, censoring at last available HIV-1 RNA before treatment discontinuation. We assessed the relative hazard/importance of GSSs according to distinct interpretation systems (Rega, ANRS and HIVdb) and other covariates by means of Cox regression and random survival forests (RSF). Prediction models were validated via the bootstrap and c-index measure.

Results

The dataset included 2337 regimens from 2182 patients, of which 733 were previously treatment-naïve. We observed 1067 virologic failures over 2820 persons-years. Multivariable analysis revealed that low GSSs of cART were independently associated with the hazard of a virologic failure, along with several other covariates. Evaluation of predictive performance yielded a modest ability of the Cox regression to predict the virologic endpoint (c-index≈0.70), while RSF showed a better performance (c-index≈0.73, p < 0.0001 vs. Cox regression). Variable importance according to RSF was concordant with the Cox hazards.

Conclusions

GSSs of cART and several other covariates were investigated using linear and non-linear survival analysis. RSF models are a promising approach for the development of a reliable system that predicts time to virologic failure better than Cox regression. Such models might represent a significant improvement over the current methods for monitoring and optimization of cART.

Machine learning study for the prediction of transdermal peptide

In order to develop a computational method to rapidly evaluate transdermal peptides, we report approaches for predicting the transdermal activity of peptides on the basis of peptide sequence information using Artificial Neural Network (ANN), Partial Least Squares (PLS) and Support Vector Machine (SVM). We identified 269 transdermal peptides by the phage display technique and use them as the positive controls to develop and test machine learning models. Combinations of three descriptors with neural network architectures, the number of latent variables and the kernel functions are tried in training to make appropriate predictions. The capacity of models is evaluated by means of statistical indicators including sensitivity, specificity, and the area under the receiver operating characteristic curve (ROC score). In the ROC score-based comparison, three methods proved capable of providing a reasonable prediction of transdermal peptide. The best result is obtained by SVM model with a radial basis function and VHSE descriptors. The results indicate that it is possible to discriminate between transdermal peptides and random sequences using our models. We anticipate that our models will be applicable to prediction of transdermal peptide for large peptide database for facilitating efficient transdermal drug delivery through intact skin.

Antiretroviral therapy optimisation without genotype resistance testing: a perspective on treatment history based models

Background

Although genotypic resistance testing (GRT) is recommended to guide combination antiretroviral therapy (cART), funding and/or facilities to perform GRT may not be available in low to middle income countries. Since treatment history (TH) impacts response to subsequent therapy, we investigated a set of statistical learning models to optimise cART in the absence of GRT information.

Methods and Findings

The EuResist database was used to extract 8-week and 24-week treatment change episodes (TCE) with GRT and additional clinical, demographic and TH information. Random Forest (RF) classification was used to predict 8- and 24-week success, defined as undetectable HIV-1 RNA, comparing nested models including (i) GRT+TH and (ii) TH without GRT, using multiple cross-validation and area under the receiver operating characteristic curve (AUC). Virological success was achieved in 68.2% and 68.0% of TCE at 8- and 24-weeks (n = 2,831 and 2,579), respectively. RF (i) and (ii) showed comparable performances, with an average (st.dev.) AUC 0.77 (0.031) vs. 0.757 (0.035) at 8-weeks, 0.834 (0.027) vs. 0.821 (0.025) at 24-weeks. Sensitivity analyses, carried out on a data subset that included antiretroviral regimens commonly used in low to middle income countries, confirmed our findings. Training on subtype B and validation on non-B isolates resulted in a decline of performance for models (i) and (ii).

Conclusions

Treatment history-based RF prediction models are comparable to GRT-based for classification of virological outcome. These results may be relevant for therapy optimisation in areas where availability of GRT is limited. Further investigations are required in order to account for different demographics, subtypes and different therapy switching strategies.

Automatic diagnosis of vocal fold paresis by employing phonovibrogram features and machine learning methods

The clinical diagnosis of voice disorders is based on examination of the rapidly moving vocal folds during phonation (f0: 80-300Hz) with state-of-the-art endoscopic high-speed cameras. Commonly, analysis is performed in a subjective and time-consuming manner via slow-motion video playback and exhibits low inter- and intra-rater reliability. In this study an objective method to overcome this drawback is presented being based on Phonovibrography, a novel image analysis technique. For a collective of 45 normophonic and paralytic voices the laryngeal dynamics were captured by specialized Phonovibrogram features and analyzed with different machine learning algorithms. Classification accuracies reached 93% for 2-class and 73% for 3-class discrimination. The results were validated by subjective expert ratings given the same diagnostic criteria. The automatic Phonovibrogram analysis approach exceeded the experienced raters' classifications by 9%. The presented method holds a lot of potential for providing reliable vocal fold diagnosis support in the future.

Prediction of prognostic biomarkers for interferon-based therapy to hepatitis C virus patients: a meta-analysis of the NS5A protein in subtypes 1a, 1b, and 3a

Background

Hepatitis C virus (HCV) is a worldwide health problem with no vaccine and the only approved therapy is Interferon-based plus Ribavarin. Response prediction to treatment has health and economic impacts, and is a multi-factorial problem including both host and viral factors (e.g: age, sex, ethnicity, pre-treatment viral load, and dynamics of the HCV non-structural protein NS5A quasispecies). We implement a novel approach for extracting features including informative markers from mutations in the non-structural 5A protein (NS5A), specifically its Interferon sensitivity determining region (ISDR) and V3 regions, and use a novel bioinformatics approach for pattern recognition on the NS5A protein and its motifs to find biomarkers for response prediction using class association rules and comparing the predictability of the different features.

Results

A total of 58 sequences from sustained responders and 94 from non-responders were downloaded from the HCV LANL database. Site-specific signatures for response prediction from the NS5A protein were extracted from the alignments. Class association rules were generated (e.g.: sustained response is associated with position A2368T in subtype 1a (support 100% and confidence 52.19%); in subtype 1b, response is associated with E2356G/D/K (support 76.3% and confidence 67.3%).

Conclusion

The V3 region was a more accurate biomarker than the ISDR region. Subtype-specific class association rules gave better support and confidence than profile hidden Markov models HMMs scores, genetic distances or number of variable sites, and would thus aid in the prediction of prognostic biomarkers and improve the accuracy of prognosis. Sites-specific class association rules in the V3 region of the NS5A protein have given the best support and confidence.

Modelling response to HIV therapy without a genotype: an argument for viral load monitoring in resource-limited settings

In the absence of widespread access to individualized laboratory monitoring, which forms an integral part of HIV patient management in resource-rich settings, the roll-out of highly active antiretroviral therapy (HAART) in resource-limited settings has adopted a public health approach based on standard HAART protocols and clinical/immunological definitions of therapy failure. The cost-effectiveness of HIV-1 viral load monitoring at the individual level in such settings has been debated, and questions remain over the long-term and population-level impact of managing HAART without it. Computational models that accurately predict virological response to HAART using baseline data including CD4 count, viral load and genotypic resistance profile, as developed by the Resistance Database Initiative, have significant potential as an aid to treatment selection and optimization. Recently developed models have shown good predictive performance without the need for genotypic data, with viral load emerging as by far the most important variable. This finding provides further, indirect support for the use of viral load monitoring for the long-term optimization of HAART in resource-limited settings.