A Tandem Evolutionary Algorithm for Identifying Causal Rules from Complex Data

Identifying neuroanatomical and behavioral features for autism spectrum disorder diagnosis in children using machine learning

Autism spectrum disorder (ASD) is a neurodevelopmental disorder that can cause significant social, communication, and behavioral challenges. Diagnosis of ASD is complicated and there is an urgent need to identify ASD-associated biomarkers and features to help automate diagnostics and develop predictive ASD models. The present study adopts a novel evolutionary algorithm, the conjunctive clause evolutionary algorithm (CCEA), to select features most significant for distinguishing individuals with and without ASD, and is able to accommodate datasets having a small number of samples with a large number of feature measurements. The dataset is unique and comprises both behavioral and neuroimaging measurements from a total of 28 children from 7 to 14 years old. Potential biomarker candidates identified include brain volume, area, cortical thickness, and mean curvature in specific regions around the cingulate cortex, frontal cortex, and temporal-parietal junction, as well as behavioral features associated with theory of mind. A separate machine learning classifier (i.e., k-nearest neighbors algorithm) was used to validate the CCEA feature selection and for ASD prediction. Study findings demonstrate how machine learning tools might help move the needle on improving diagnostic and predictive models of ASD.

Survey of awareness, attitudes, and compliance with COVID-19 measures among Vermont residents

The objective of this research was to examine residents’ awareness, attitudes, and compliance with COVID-19 public health guidelines in Vermont, which emerged as an early leader in national pandemic response. Our methods included conducting an online survey of adult Vermont residents between January and April 2021. We analyzed demographics associated with awareness and compliance, and identified features associated with non-compliance. Our results show that of the 2,208 adult Vermont residents who completed the survey, 90% were extremely aware of the state’s COVID-19 guidelines, and 95% reported knowing exactly what to do to follow recommended actions. Political affiliation emerged as a primary factor related to attitudes and compliance. Self-identified Republicans were less likely to agree that public health measures keep people safe or help businesses stay open, and were less likely to follow masking, quarantine, social distancing, and vaccine guidance than Independents, Progressives, and Democrats. The large differences in COVID-19 infection and death rates across the country, and recent shift toward a "pandemic of the unvaccinated," underscore the need for identifying public health strategies that work in some areas in order to adapt and apply them to areas that have struggled with controlling the virus. Consistent with national surveys, our results show that resistance to public health guidance is a partisan challenge even in states with high compliance. Identifying populations that are less supportive or hesitant to follow guidelines while understanding factors that motivate compliance can help inform strategies for developing targeted programs to encourage collective action on pandemic response. Developing communication strategies that reach people who do not believe COVID-19 guidelines keep them safe is necessary to reach universal compliance.

Use of Artificial Intelligence and Machine Learning for Discovery of Drugs for Neglected Tropical Diseases

Neglected tropical diseases continue to create high levels of morbidity and mortality in a sizeable fraction of the world’s population, despite ongoing research into new treatments. Some of the most important technological developments that have accelerated drug discovery for diseases of affluent countries have not flowed down to neglected tropical disease drug discovery. Pharmaceutical development business models, cost of developing new drug treatments and subsequent costs to patients, and accessibility of technologies to scientists in most of the affected countries are some of the reasons for this low uptake and slow development relative to that for common diseases in developed countries. Computational methods are starting to make significant inroads into discovery of drugs for neglected tropical diseases due to the increasing availability of large databases that can be used to train ML models, increasing accuracy of these methods, lower entry barrier for researchers, and widespread availability of public domain machine learning codes. Here, the application of artificial intelligence, largely the subset called machine learning, to modelling and prediction of biological activities and discovery of new drugs for neglected tropical diseases is summarized. The pathways for the development of machine learning methods in the short to medium term and the use of other artificial intelligence methods for drug discovery is discussed. The current roadblocks to, and likely impacts of, synergistic new technological developments on the use of ML methods for neglected tropical disease drug discovery in the future are also discussed.

Novel Evolutionary Algorithm Identifies Interactions Driving Infestation of Triatoma dimidiata, a Chagas Disease Vector

Chagas disease is a lethal, neglected tropical disease. Unfortunately, aggressive insecticide-spraying campaigns have not been able to eliminate domestic infestation of Triatoma dimidiata, the native vector in Guatemala. To target interventions toward houses most at risk of infestation, comprehensive socioeconomic and entomologic surveys were conducted in two towns in Jutiapa, Guatemala. Given the exhaustively large search space associated with combinations of risk factors, traditional statistics are limited in their ability to discover risk factor interactions. Two recently developed statistical evolutionary algorithms, specifically designed to accommodate risk factor interactions and heterogeneity, were applied to this large combinatorial search space and used in tandem to identify sets of risk factor combinations associated with infestation. The optimal model includes 10 risk factors in what is known as a third-order disjunctive normal form (i.e., infested households have chicken coops AND deteriorated bedroom walls OR an accumulation of objects AND dirt floors AND total number of occupants ≥ 5 AND years of electricity ≥ 5 OR poor hygienic condition ratings AND adobe walls AND deteriorated walls AND dogs). Houses with dirt floors and deteriorated walls have been reported previously as risk factors and align well with factors currently targeted by Ecohealth interventions to minimize infestation. However, the tandem evolutionary algorithms also identified two new socioeconomic risk factors (i.e., households having many occupants and years of electricity ≥ 5). Identifying key risk factors may help with the development of new Ecohealth interventions and/or reduce the survey time needed to identify houses most at risk.