As part of a translational research team within the Department of Neurological Surgery, I'm contributing to multiomic-based research efforts to further the groups initiative in brain cancer research and treatment.

As part of an NIH SBIR funded research project, I collaborated with a team at Seattle Children's Research Institute (SCRI) to contribute to a translational effort to advance PreCyte's blood-based diagnostic assay, iCAP. Initially developed at the Institute for Systems Biology (ISB), this assay uses cells as biosensors to monitor gene expression changes in response to environmental queues.I led biomarker discovery and ML modeling/prediction initiatives within a retrospective clinical study aimed at early lung cancer detection. In addition to analytical contributions, I also worked to develop an infrastructure roadmap to support AI/ML classifier development at scale.

My primary focus in this role was to investigate the diagnostic potential of the iCAP assay for early detection of Alzheimer's disease (AD). Using cells as biosensors, I employed RNA-Seq and network-based approaches to study gene expression changes in iPSC-derived neurons, aiming to identify potential biomarkers capable of discriminating between pre-MCI, MCI, non-AD, and normal cohort members.I applied computational biology approaches alongside a comprehensive machine learning workflow encompassing batch correction, outlier removal, confounder identification, feature selection, and model training. Utilized logistic regression, glmnet, tree-based methods, and boosting approaches. Performed error estimation and prediction tasks to assess diagnostic performance.Additionally, I proposed and conducted analyses that explored the emerging infection hypothesis in AD progression. Utilizing bulk RNA-Seq data, I examined the correlation between AD-related iCAP signatures and the presence of low levels of viral and bacterial sequence content in plasma, identified through the detection and characterization of 16S rRNA sequences.I developed AWS-based analysis workflows, enabling end-to-end analysis of RNA-Seq data at scale. As an advocate for reproducible research, I established best practices by creating analysis workflows and methodologies and automated build processes for shared Docker images.

Expertise in: - Machine Learning - Statistical Analyses - Genomic Analysis (NGS) - Gene Expression Analysis - Data Visualization - Python and R - Scalable Analytics

AlloVir is a gene therapy startup (Series B) focused on developing viral therapies for patients having a weakened immune system. I am exploring relationships between primary and secondary endpoints in Phase 3 clinical trial data, distilling complex trends from an extensive clinical trial database. Following AlloVir's strategic decision to discontinue their global Phase 3 trial, I will be helping to explore their unblinded data.This work is under contract via BioBridges.

Working closely with clinical research scientists, I utilized machine learning techniques to evaluate and improve the accuracy of somatic variant calls in NGS data obtained from FFPE samples. In a separate project, I took the lead in a collaborative effort with the Engineering team, where we developed a custom visualization tool for analyzing time-series data. The tool ensured that our instruments consistently met the required performance standards. Additionally, I played an active role in Business Development by creating signal analysis algorithms for nanopore DNA sequencing data, contributing to a successful acquisition.

In close collaboration with the wet-lab and Engineering teams, I contributed to early research and development (R&D) initiatives, focusing on the enhancement of capture uniformity for SeqCap EZ target enrichment products, including whole exome and panel designs. I developed algorithms to quantify and improve coverage over challenging genomic regions, particularly those with extreme GC richness. Additionally, I evaluated the fidelity of probes generated by NimbleGen's maskless array synthesizer (MAS).Also, I actively contributed to solid-state nanopore DNA sequencing research in partnership with IBM. Within this project, I conducted DNA nanopore data analysis and software development. Notably, I successfully ported peak detection algorithms from C++ to MATLAB, increasing accessibility for the entire team. Additionally, I designed and implemented an AWS solution for executing scalable workflows for NGS data prior to the availability of more off-the-shelf solutions.