In the hypothetical scenario that a weaponized biological agent is suspected in an attack on United States interests, it is imperative that field samples be rapidly collected and transferred to specialized laboratories for analysis.
The goal of our lab at the Biological Defense Research Directorate was to employ DNA sequencing and microarray technologies to rapidly identify biological warfare agents, characterize patterns of DNA sequence and strain variation, and determine targets for development of novel vaccine and diagnostics.
To this end, I designed the analytical framework to predict and classify sequence features from raw genomic data including genes, pseudo-genes, transcriptional elements, and other non-coding RNAs using a combination of interpolated Markov models, probabilistic covariance models, local alignment algorithms, and curated databases of functional annotations. While even microbial genomes can be computationally expensive to analyze, the data pipeline utilized a distributed computing cluster for parallel processing. The analyzed data was loaded into a relational database for frequent SQL querying and data visualization.
Analytics from the collected data was used to scan and characterize hundreds of samples, improving identification processes and contributing to several studies.