Woolf, T. | bioRxiv Science

bioRxiv2021

Machine Learning to Summarize and Provide Context for Sleep and Eating Schedules.

The relative timing of sleep and of eating within the circadian day is important for human health. Despite much data on sleep, and a growing data set for eating, there remains a need for an interpretative framework for the understanding of this data for health decisions. This study provides a new statistical and machine learning analysis of more than 500 participants in the Daily24 project. From their data, and the analysis, we propose a framework for determining the classification of participants into different chronotypes and with that the ability to realize the potential impact of daily circadian habits on health. We propose that our resulting distribution curves could be used, similar to NHANES (National Health and Nutrition Examination Survey) data for pediatric growth, as a measure for circadian misalignment and used to help guide re-entrainment schedules. Author summaryDaily habits can be positive, negative or neutral for human health. Generally sleep and eating schedules are assumed without thought for their potential to help or interfere with health. In this study we propose a framework, based on data from more than 500 participants, for evaluating the relative timing of meals and sleep schedules. This evaluation, similar to pediatric growth charts, can guide clinical suggestions for those at the extremes, while helping others to realize that they are unusual relative to the population average

bioinformatics↗

bioRxiv2019

Feature Design for Protein Interface hotspots using KFC2 and Rosetta

Protein-protein interactions regulate many essential biological processes and play an important role in health and disease. The process of experimentally charac-terizing protein residues that contribute the most to protein-protein interaction affin-ity and specificity is laborious. Thus, developing models that accurately characterize hotspots at protein-protein interfaces provides important information about how to inhibit therapeutically relevant protein-protein interactions. During the course of the ICERM WiSDM workshop 2017, we combined the KFC2a protein-protein interaction hotspot prediction features with Rosetta scoring function terms and interface filter metrics. A 2-way and 3-way forward selection strategy was employed to train support vector machine classifiers, as was a reverse feature elimination strategy. From these results, we identified subsets of KFC2a and Rosetta combined features that show improved performance over KFC2a features alone.

bioinformatics↗