Exemplary Projects 

Investigators: Dean Regier (lead), Anita Palepu, Deirdre Weymann, Emanuel Krebs 

Summary: Working with patients, communities and the public, the goal of this project is to determine the acceptability of AI-augmented decision-making in patient care. Researchers will build a custom natural language processing (NLP) pipeline trained to effectively extract topics of discussion and classify sentiment from a variety of different natural language sources (e.g., focus group and townhall transcripts, online discussion forums) across healthcare domains.  

The AI-enabled platform will perform longitudinal and rapid synthesis of online discourse, interactions and changes in sentiment over time. These inputs will help determine facilitators and barriers of routine implementation of AI by the bedside into the B.C. healthcare system, with the goal of ensuring AI is introduced in a way that aligns with patient values and expectations. 

Investigators: Stephen Lam (co-lead), Raymond Ng (co-lead), Leonid Sigal, Lovedeep Gondara 

Summary: Working closely with the B.C. Ministry of Health, Provincial Health Services Authority (PHSA) and breast and lung cancer oncologists at BC Cancer, the goal is to expedite triage from cancer screening to provide timely access to curative therapies and improve long-term survival of patients. The project team has already developed an NLP tool that flags high-risk breast cancer patients from mammography for timely access to treatments. In parallel, the team will expand its work to include early detection and screening of lung cancer, where they will develop NLP tools to identify incidental lung nodules from imaging reports and develop segmentation tools to map identified lung nodules onto the actual images.

Investigators: Daniel Vigo (co-lead), Raymond Ng (co-lead) 

Summary: The goal of this project is to implement screening and intervention tools to support student mental health using an existing AI-enhanced mobile mental health app called Minder. A retrieval-augmented generation large language model (RAG-LLM) will be developed based on expert documentation on mental health and substance use. This will be used to perform mood detection on user text input data and will ensure there is accurate information based on the source documents. If the system detects a “signal of concern” that the student is struggling, early intervention will provide them with appropriate and safe responses from a carefully curated library of expert-human-generated responses. This project will serve as a major demonstration project on how to realize the potential of AI by safely integrating it into an existing digital tool embedded within the system of mental health services by working with PHSA.   

Investigators: Jörg Gsponer (lead), Jennifer Bui, Filip Van Petegem 

Summary: Many cardiac disorders, such as arrhythmias, are linked to mutations in calcium channel regulatory sites—regions that are often intrinsically disordered and difficult to characterize or target therapeutically. Recent advances in AI-guided protein design offer promising strategies to overcome these challenges by enabling the development of high-affinity binders or engineered regulatory partner proteins. This project aims to test the computational design of such proteins to modulate calcium channel function and correct disease-related deficits. As a proof of concept, a successful outcome would highlight the transformative potential of AI in creating novel therapeutics. 

Investigators: Sriram Subramaniam (lead) 

Summary: Resistance to drug treatments emerges in both cancer and viral infections, and each time this happens, present day approaches require starting the lengthy and expensive process from scratch to develop new drugs. This project aims to address this problem by combining atomic resolution imaging using cryogenic electron microscopy (cryo-EM) with the latest advances in protein design and generative AI to rapidly engineer, test, develop and deliver antibody therapies that bypass the traditional process of antibody discovery. The project’s initial focus will be the application of AI-driven design to convert an existing, but no longer effective, SARS-CoV-2 antibody to become effective against a new variant. The project team will then apply this in-house AI workflow to design AI-derived antibodies that target yet-to-emerge SARS-CoV-2 variants, and other viral and cancer targets. 

Investigators: Ali Bashashati (lead) 

Summary: AI and deep learning are accelerating traditional drug discovery through computational screening of potential compounds. However, frequent false predictions highlight the need for improved algorithms that can refine their accuracy based on experimental feedback.  

A key challenge is the lack of high-quality, real-world data—known as ground truth data—for deep learning algorithm development, refinement and testing. When scaled, the data generated through this initiative could become a substantial resource to refine algorithms and train foundation models for this specific problem. 

Focusing on cancer immunotherapies, specifically immune checkpoint inhibitors, the goal of this project is to leverage recent breakthroughs in AI-driven computational compound screening and high-throughput structural biology by incorporating highly efficient biosensors into the drug development pipeline. These sensors will be used to generate massive amounts of experimental ground truth data that could be used for designing new and more accurate AI-based drug discovery algorithms.  

Investigators: Dean Regier (lead), Sriram Subramaniam, Ali Bashashati, Anna Blakney 

Summary: Called the “Valley of Death,” fewer than 14 per cent of innovative health products reach patients. For those that do, the journey to achieve regulatory approval, reimbursement, and implementation takes 17 to 24 years. 

This project aims to change that. Its goal is to create a generative AI regulatory science platform that empowers UBC and B.C. biotech researchers to navigate translational roadblocks in parallel with the scientific discovery process. Researchers will establish an LLM that generates tailored guidance and reports on competitor technologies and their stage of development (landscape analyses), accelerated commercialization, regulatory and reimbursement pathways, and equity-focused implementation to bring innovative products to patients more quickly and equitably—both in B.C. and around the world.