Ensemble machine learning modeling for the prediction of artemisinin resistance in malaria

Abstract

Resistance in malaria is a growing concern affecting many areas of Sub-Saharan Africa and Southeast Asia. Since the emergence of artemisinin resistance in the late 2000s in Cambodia, research into the underlying mechanisms has been underway. The 2019 Malaria Challenge posited the task of developing computational models that address important problems in advancing the fight against malaria. The first goal was to accurately predict artemisinin drug resistance levels of Plasmodium falciparum isolates, as quantified by the IC50. The second goal was to predict the parasite clearance rate of malaria parasite isolates based on in vitro transcriptional profiles. In this work, we develop machine learning models using novel methods for transforming isolate data and handling the tens of thousands of variables that result from these data transformation exercises. This is demonstrated by using massively parallel processing of the data vectorization for use in scalable machine learning. In addition, we show the utility of ensemble machine learning modeling for highly effective predictions of both goals of this challenge. This is demonstrated by the use of multiple machine learning algorithms combined with various scaling and normalization preprocessing steps. Then, using a voting ensemble, multiple models are combined to generate a final model prediction.

Machine learning for artemisinin resistance in malaria treatment across in vivo-in vitro platforms

“…This is a meaningful step in the research of malaria treatment, as the work demonstrated the potential and robustness of a personalized model for ART resistance, which has not been achieved before. Some studies addressed the prediction on either in vivo or in vitro study but did not generalize the model across different conditions ( Ford and Janies, 2020 ; Li et al., 2021 ; Sastry et al., 2021 ). In fact, previous studies reported that generating predictive models for ART resistance has been challenging, as the in vitro IC50 of P. falciparum in standard drug susceptibility assay correlates poorly with its clearance rate in vivo ( Chotivanich et al., 2014 ; Fairhurst and Dondorp, 2016 ).…”

Modeling Plasmodium falciparum Diagnostic Test Sensitivity using Machine Learning with Histidine-Rich Protein 2 Variants

“…Beyond simply training a better model using more sophisticated algorithms, our research focus is to allow for interpretable insights of the machine learning models to be derived from the "black box". We have shown previous success in AI-driven explanations of gene expression underlying drug resistant strains of Plasmodium falciparum [6,7]. We apply this model interpretability here to identify which types of histidine-rich repeats, present in PfHRP2, are most indicative of malaria test performance.…”