Speaker
Description
If two lineages differ by a single mutation, can wastewater lineage abundance estimators (such as Freyja) actually distinguish them? Can the presence of mutations be predicted from previous wastewater samples, possibly with machine learning methods? Can we use clustering algorithms to detect new lineages from wastewater without prior knowledge of any lineage definitions?
These questions are all different aspects of "discovery": either discovering the presence of known lineages or the discovery of new lineages based on wastewater alone. The answer to all these questions is an emphatic "yes!", followed by an emphatic "and we get to use fun statistical theory to quantify our uncertainties!" This work presents a deep dive into the challenges (also known as "opportunities") we face when using a list of mutations to discover lineages in wastewater.
