Unearthing the Dark Antigenome & Leveraging Machine Learning to Accurately Prioritize Neoantigens
Time: 9:05 am
day: Pre-Conference Day
Details:
High throughput sequencing technologies are allowing for an ever-increasing resolution of tumour genome characterization, offering exciting windows of opportunity for (personalized) cancer vaccines. How should we identify the most promising targets, with the highest probability of (T cell) immunogenicity? Are there new targets in the genome that were previously left not considered? What roles can machine learning and AI play in prioritizing candidates among the plethora of choices that are available for most patients with cancer? What are key considerations in applying these technologies correctly?
In this workshop we will discuss our efforts in both antigen discovery and computational validation of immunogenicity at CureVac. During presentations, we’ll identify controversial topics and call for audience input.
Discovery of Unannotated Small Open Reading Frame Tumour Associated Antigens
- smORF TAA antigen discovery
- Decreasing search spaces, assessing translational potential
- Experimental validation
Machine learning for Neoantigen Identification
- Predictive features to consider in immunogenicity prediction
- Robust machine learning to learn the characteristics of immunogenic neo-epitopes
- Limitations to the currently available training data and how to address them
