Publications and Preprints

BioKlustering: a web app for semi-supervised learning of maximally imbalanced genomic data

Summary: Accurate phenotype prediction from genomic sequences is a highly coveted task in biological and medical research. While machine-learning holds the key to accurate prediction in a variety of fields, the complexity of biological data can render many methodologies inapplicable. We introduce BioKlustering, a user-friendly open-source and publicly available web app for unsupervised and semi-supervised learning specialized for cases when sequence alignment and/or experimental phenotyping of all classes are not possible. Among its main advantages, BioKlustering 1) allows for maximally imbalanced settings of partially observed labels including cases when only one class is observed, which is currently prohibited in most semi-supervised methods, 2) takes unaligned sequences as input and thus, allows learning for widely diverse sequences (impossible to align) such as virus and bacteria, 3) is easy to use for anyone with little or no programming expertise, and 4) works well with small sample sizes. Availability and Implementation: BioKlustering this https URL is a freely available web app implemented with Django, a Python-based framework, with all major browsers supported. The web app does not need any installation, and it is publicly available and open-source this https URL.

Download here

Identifying microbial drivers in biological phenotypes with a Bayesian Network Regression model

Published in Ecology and Evolution, 1900

Abstract: In Bayesian Network Regression models, networks are considered the predictors of continuous responses. These models have been successfully used in brain research to identify regions in the brain that are associated with specific human traits, yet their potential to elucidate microbial drivers in biological phenotypes for microbiome research remains unknown. In particular, microbial networks are challenging due to their high-dimension and high sparsity compared to brain networks. Furthermore, unlike in brain connectome research, in microbiome research, it is usually expected that the presence of microbes have an effect on the response (main effects), not just the interactions. Here, we develop the first thorough investigation of whether Bayesian Network Regression models are suitable for microbial datasets on a variety of synthetic data that was generated under realistic biological scenarios. We test whether the Bayesian Network Regression model that accounts only for interaction effects (edges in the network) is able to identify key drivers in phenotypic variability (microbes). We show that this model is indeed able to identify influential nodes and edges in the microbial networks that drive changes in the phenotype for most biological settings, but we also identify scenarios where this method performs poorly which allows us to provide practical advice for domain scientists aiming to apply these tools to their datasets. Finally, we implement the model in a Julia package available via Github or the Julia package repository.

Download here