Jeremy Wang: Classification of pediatric acute leukemia using full-length transcriptomics
Jeremy Wang started the Human & Translational Research track off with his talk on pediatric acute leukemia. As a background to his work, there is a clear global discrepency in in childhood cancer 5-year survival rates, with a glaringly obvious pattern of poorer prognosis in lower income countries, includingparts of Africa, Asia, and America. This is underpinned by capital cost of equipment for characterising and diagnosis, particularly within genomics. Low-cost nanopore sequencing can leapfrog the barrier set by incumbent technologies.
Gene expression profiling alone has been demonstrated able to classify paediatric leukaemiaโs and their lineages, historically conducted with microarrays. This was an inflexible aproach, leaving much to be desired, and while short-read RNA sequencing is robust it carries significant capital costs. Jeremy set out to demonstrate that comparable classification of paediatric leukemias was possilble in low-resource settings using nanopore sequencing. This he demonstrated, even when using significantly degraded RNA โwhere many shorter reads exist, their sampling density across the transcriptome was comparable or higher than traditional short-read methods.
In order to achieve this he developed a number of machine learning and bioinformatics tools, combining partial least-squares (PLS) regression and a support vector machine (SVM) design to distinguish between expression profiles. Work with 211 nanopore-based transcriptomics profiles demonstrated promising results of almost 98% accuracy in aggregate, allowing him to define conserative cut-off values which accurately classify 95% of samples. For future work Jeremy plans to further improve this gene expression-based classificaiton, but importantly he is working with international collaborators in Malawi, Pakistan,and El Salvado to assess the feasibility of deploying his pipeline.
No comments found