October 3 – Jessica (Jingyi) Li, Department of Statistics, UCLA
Abstract: The rapid development of genomics technologies has propelled fast advances in genomics data science. While new computational algorithms have been continuously developed to address cutting-edge biomedical questions, a critical but largely overlooked aspect is the statistical rigor. In this talk, I will introduce our recent work that aims to enhance statistical rigor by addressing three issues:
- Large-scale feature screening (i.e., enrichment and differential analysis of high-throughput data) relying on ill-posed p-values;
- Double-dipping (i.e., statistical inference on biasedly altered data);
- Gaps between black-box generative models and statistical inference.
October 10 – Luca Mazzucato, Department of Mathematics and Biology, University of Oregon
Abstract: Animal behavior exhibits a striking amount of variability in the temporal domain along at least three independent axes: hierarchical, contextual, and stochastic. First, a vast hierarchy of timescales links movements into behavioral sequences and long-term activities, from milliseconds to minutes. Second, action timing can be modulated by changes in context, of either internal (neuromodulatory, state-dependent) or external origin. Third, self-initiated actions exhibit large residual variability across repetitions, with signatures of stochastic origin. What computational principles underlie such complex temporal features? We will present the foundation of a theory of temporal variability in behavior and neural activity, based on metastable attractors observed in sensory and motor cortical areas. We will highlight the essential role played by intrinsic noise and heterogeneities in controlling the features of temporal variability.
October 17 – Hao Chen, Department of Statistics, UC Davis
Abstract: After observing snapshots of a network, can we tell if there has been a change in dynamics? After collecting spiking activities of thousands of neurons in the brain, how shall we extract meaningful information from the recording? We introduce a change-point analysis framework utilizing graphs representing the similarity among observations. This approach is non-parametric and can be applied to data when an informative similarity measure can be defined. Analytic approximations to the significance of the test statistics are derived to make the method fast applicable to long sequences. The method is illustrated through the analysis of the Neuropixels data.
October 24 – Ali Shojaie, Department of Biostatistics, University of Washington, Seattle
Abstract: Recent evidence suggests that changes in biological networks, e.g., rewiring or disruption of key interactions, may be associated with development of complex diseases. These findings have motivated new research in computational and experimental biology that aim to obtain condition-specific estimates of biological networks, e.g. for normal and tumor samples, and identify differential patterns of connectivity in such networks, known as differential network analysis. In this talk, we primarily focus on testing whether two Gaussian graphical models are the same. We will first illustrate that existing inference procedures for this task may lead to misleading results. To address this shortcoming, we propose a two-step inference framework, for testing the null hypothesis that the edge sets in two networks are the same. The proposed framework is especially appropriate if the goal is to identify nodes or edges that show differential connectivity. Time permitting, we will also discuss how differential network analysis methods can be extended to non-Gaussian settings as well as settings where differences in network edges are functions of other covariates.
October 31 – Dominik Rothenhausler, Department of Statistics, Stanford University
November 7 – Karthika Mohan, EECS Department, Oregon State University
November 14 – Rahul Majumder, Operations Research and Statistics Group, Sloan School of Business, MIT
November 21 – Vijayan Nair, Wells Fargo
November 28 – Ben Brown, Department of Statistics, UC Berkeley, and Computational Biologist, Berkeley Labs