On 12 September, Dr Qinghua Song, who is an Associate Director at Biostatistics Gilead, gave a talk at UIC titled, “Applications of Machine Learning Methods in Biomarker Analyses in Clinical Development”. This talk was organised by the UIC Statistics programme and was part of the Division of Science and Technology (DST) lecture series.

2Q2A0170
Dr Song is the Associate Director at Biostatistics Gilead

Dr Song began his talk by introducing a little bit about himself, such as that he got his PhD from the Department of Statistics, University of Wisconsin-Madison in 2005. Since then he has worked in world renowned pharmaceutical companies such as Merck, Genentech and Gilead.

2Q2A0126

He then gave a brief overview of drug development from a statistician’s point of view as well as introducing biomarkers in drug discovery and development. He mostly spoke about biomarker, which in medicine is a measurable indicator of the severity or presence of some disease state.

Using his wealth of knowledge and experience, Dr Song discussed how statisticians are involved in the design, power calculation, data management and analysis of studies, communication of results and report writing in drug discovery and development.


Dr Song explaining about the important roles of statisticians in drug development

During the talk, Dr Song used examples of his past experience where he provided statistical analysis support for research. Also it was worth noting that he co-authored on multiple scientific papers in application of machine learning, biomarker selection, application of statistical modeling and analysis on large data.

His experiences include managing a group of statisticians for the early phase virology studies in Gilead Science. He is leading a group of data scientists, to develop efficient and innovative statistical tools for data visualisation and advanced analytical methods.

2Q2A0184

The lecture generally discussed two case studies about applying machine learning methods for biomarker analyses. Case 1 was about ‘Application of Random Forest to building Proteomics Classifier’ while Case 2 was about ‘Gene signature derivation using SuperLearning with application to the epithelial mesenchymal transition in lung cancer’.

Reporters/Photographers: Samuel Burgess, Marissa Furney
Editors: Deen He, Étienne Fermie
(from MPRO)