Lu contributes to algorithm, identifies threats to humans, crops

Jacob Lu's work in the Treangen Lab

Lu contributes to algorithm, identifies threats to humans, crops

Jacob Lu (CS ’22) contributed to an algorithm to predict DNA sequences that are harmful to humans, crops, and other organisms.

Lu is part of assistant professor Todd Treangen’s Computational Forensics and Metagenomics research group.

“It was a wonderful to have Jacob in the lab. He made substantial contributions to the project that helped improve the performance of software that rapidly and robustly screens DNA for threats,” Treangen said.

Lu was interested in seeing his research come to life.

“Professor Treangen leads a high-energy lab driven by his ambition and passion for genomics. I decided to work with him because I sensed I’d be held accountable to make real contributions,” he said.

“My previous experience with machine learning also equipped me to jump into one of the lab’s fastest-paced projects: a government assignment with impacts on public safety. The purpose behind our research—to make DNA synthesis safer—motivated me during my internship,” Lu said.

Treangen said the goal of this project is to develop a robust software solution for assessing the threat potential of short DNA sequences targeted for rapid yet sensitive screening of oligonucleotides. This project is now in its second phase, and his team is one of the three computational performers on the Functional Genomic and Computational Assessment of Threats (Fun GCAT) program, managed by IARPA within the office of the director of National Intelligence.

“Specifically, we aspire to enhance DNA screening of sequences that might accidentally, or intentionally, be altered to result in a synthetic biological threat,” Treangen said.

“Our team designed software to screen short DNA sequences and assess for potentially harmful function,” Lu said.

“I developed machine learning algorithms to predict which sequences would be harmful to humans, crops, and other organisms. I also created a graph-database to store the machine learning predictions and visualize protein relationships,” he said.

Lu worked in Dr. Bruce Shapiro's RNA structural biology lab at the NIH in Frederick, Maryland, where he developed neural networks to predict RNA transcriptional pause sites, worked on simulating RNA and DNA folding, and created RiboSketch, a tool to produce RNA and DNA secondary-structure diagrams.

This work is in collaboration with Signature Science LLC, the Fraunhofer Center for Experimental Software Engineering, and the Chris Jermaine Lab.

Cintia Listenbee, Communications and Marketing Specialist in Computer Science