Introduction

Speaker 1: Professor Gareth Roberts FRS, University of Warwick

Speaker 2: Professor Bin Yu, University of California, Berkeley

Speaker Host: Sofia Olhede, UCL

About the event

Turing Lecture banner

Speaker 1: Professor Gareth Roberts FRS, University of Warwick

New challenges in Computational Statistics

The presentation will investigate some challenges and opportunities for modern Computational Statistics within the emerging super-discipline of Data Science. Within the last 30 years, Computational Statistics has been tremendously successful in transporting theoretically well-underpinned statistical paradigms through to methodology for increasingly complex and high-dimensional statistical models; and ultimately to diverse inferential problems in science and beyond. This applies both to frequentist and Bayesian approaches. However, apparently paradoxically, statistical theory tells us that with large data sets, inference ought to be increasingly straightforward, while necessary calculations are increasingly computationally expensive. On the other hand, these difficulties, together with an increased awareness of computational hardware is inspiring new directions for the subject. The lecture will conclude with a brief description of some algorithms, which draw together crucial tools from statistics, probability, numerical analysis and computing to introduce new algorithms tackle challenging emerging inference problems emerging from data science.

Speaker 2: Professor Bin Yu, University of California, Berkeley

Unveiling the Mysteries in Spatial Gene Expression

Genome-wide data reveal an intricate landscape where gene activities are highly differentiated across diverse spatial areas. These gene actions and interactions play a critical role in the development and function of both normal and abnormal tissues. As a result, understanding spatial heterogeneity of gene networks is key to developing treatments for human diseases. Despite the abundance of recent spatial gene expression data, extracting meaningful information remains a challenge for local gene interaction discoveries. In response, we have developed staNMF, a method that combines a powerful unsupervised learning algorithm, nonnegative matrix factorization (NMF), with a new stability criterion that selects the size of the dictionary.

Using staNMF, we generate biologically meaningful Principle Patterns (PP), which provide a novel and concise representation of Drosophila embryonic spatial expression patterns that correspond to pre-organ areas of the developing embryo. Furthermore, we show how this new representation can be used to automatically predict manual annotations, categorize gene expression patterns, and reconstruct the local gap gene network with high accuracy. Finally, we discuss on-going crispr/cas9 knock-out experiments on Drosophila to verify predicted local gene-gene interactions involving gap-genes.

Speakers

Professor Bin Yu

Scientific Advisory Board Member and Chancellor's Professor in the Departments of Statistics and of Electrical Engineering & Computer Sciences at the University of California, Berkeley

Location

University of London

London, UK

51.5229378, -0.13082059999999