As a researcher with experiences in a variety of fields, Brian is a believer in the innovation and insights that can be gained from highly inter-disciplinary research. In 2011, he earned his ScB in Cognitive Neuroscience from Brown University (Providence, RI). The core focus of his work was and continues to be the investigation of human susceptibility to diseases and disorders in order to impact human health and alleviate suffering.
Brian then took on research to uncover the evolutionary underpinnings of human-specific cognitive abilities and neurobiological susceptibilities at The George Washington University (Washington, DC). At the same institution, he went on to earn an MPhil in Human Paleobiology (2017) investigating the neuroanatomical, transcriptomic and genomic evolution of the hippocampus and episodic memory. During this time, he became increasingly interested in applying programming and computational techniques to expand the scale and scope of the questions that he is able to address.
In the pursuit of further advancing his skills as a computational researcher, he thereafter accepted a position as a Bioinformatician at the Icahn School of Medicine at Mount Sinai. There he helped to develop a variety of open-access software for systems biology and genomics. Alongside a multi-disciplinary team of molecular biologists, computational biologists, and physicians, Brian published work revealing cell-type-specific neurogenomic mechanisms underlying neurological conditions through the application of statistical and machine learning techniques to large-scale multi-omics and clinical data.
Brian is now pursuing my PhD in Clinical Medicine Research at Imperial College London under the advisorship of Dr Nathan Skene.
Genome-wide association studies (GWAS) have identified hundreds of genetic loci associated with neurodegenerative diseases, including Alzheimer’s disease (AD) and Parkinson’s disease (PD). However, most risk-associated variants within these loci fall within non-protein-coding regions, making functional effect predictions and target identification challenging. Furthermore, AD/PD are highly heterogeneous in their clinical presentation and are co-morbid with other disorders (e.g. cardiovascular disease, Type II diabetes) and symptoms. Given the large number of disease/trait GWAS that have been conducted (>30,000 in the OpenGWAS database), it is now possible to identify shared genetics across many traits and computationally dissect AD/PD genetic signatures into their constituent sub-traits. This will allow for far more nuanced views of these diseases that can potentially explain the heterogeneity observed across patients.
In Brian's thesis, he will apply the following Aims to AD and PD:
1) Uncover the cell-types underlying genetic risk by improving methods for GWAS cell-type deconvolution using single-cell-resolution transcriptomic (scRNA-seq) and epigenomic datasets (scATAC-seq, scCUT&Tag).
2) Apply dimensionality reduction techniques (e.g. truncated singular value decomposition, convolutional autoencoders) to many GWAS datasets at once to discover the latent structure of diseases/trait genomics, and computationally dissect the cell-types underlying each sub-trait within AD/PD.
This project aims to not only further understanding of neurodegenerative diseases, but also provide an extensible framework that can be easily applied to any other disorder or trait for which there is relevant GWAS data.