Statistics and Probability Seminars

Statistics Department Seminar presents "Novel clinical trial designs and statistical methods in the era of precision medicine"

Topic: 
Novel clinical trial designs and statistical methods in the era of precision medicine
Abstract / Description: 

We begin with FDA's Guidance for Industry on (a) Master Protocols for Efficient Clinical Trial Designs to Expedite Development of Oncology Drugs and Biologics in 2018; (b) Enrichment Strategies for Clinical Trials to Support Determination of Effectiveness of Human Drugs and Biological Products in March 2019; and (c) Adaptive Designs for Clinical Trials of Drugs and Biologics in November 2019. We then describe their biostatistical and biopharmaceutical underpinnings, focusing on recent advancements in adaptive group sequential trial designs and statistical methods for their analysis, and conclude with challenges and opportunities for statistical science in precision medicine and regulatory submission.

This is joint work with Tze Lai and Nikolas Weissmueller.

Date and Time: 
Tuesday, June 30, 2020 - 4:30pm
Venue: 
Zoom ID: 941 3461 3493 (+password to come)

Statistics Department Seminar “Identifying condition-specific patterns in large-scale genomic data”

Topic: 
Identifying condition-specific patterns in large-scale genomic data
Abstract / Description: 

Joint analyses of genomic datasets obtained in multiple different conditions are essential for understanding the biological mechanism that drives tissue-specificity and cell differentiation. But it still remains computationally challenging even when the number of conditions is moderate.

I will present CLIMB (Composite LIkelihood eMpirical Bayes), a statistical methodology which learns patterns of condition specificity present in genomic data by leveraging pairwise information. CLIMB provides a generic framework facilitating a host of downstream analyses, such as clustering genomic features sharing similar conditional-specific patterns and identifying which of these features are involved in cell fate commitment. It improves upon existing methods by boosting statistical power to identify biologically meaningful signals while retaining interpretability and computational tractability. We illustrate CLIMB's value on a CTCF ChIP-seq dataset measured in 17 different cell populations and an RNA-seq dataset measured in three committed hematopoietic lineages. These analyses demonstrate that CLIMB captures biologically relevant clusters in the data and improves upon commonly-used pairwise comparisons and unsupervised clusterings typical of genomic analyses.

Date and Time: 
Tuesday, June 23, 2020 - 4:30pm
Venue: 
Zoom ID 941 3461 3493 (+password)

SPECIAL EVENT Probability Seminar presents "Universality of approximate message passing algorithms"

Topic: 
Universality of approximate message passing algorithms
Abstract / Description: 

Approximate Message Passing (AMP) algorithms are non-linear power iterations originally arising from the context of compressed sensing. In this talk, I will introduce a Lipschitzian functional iteration, as a generalization of the AMP algorithms, and discuss its universality in disorder. In addition, I will explain how our results imply universality in a number of AMPs popularly adapted in Bayesian inferences and optimizations in spin glasses.

This is based on a joint work with Wai-Kit Lam.


Joint with Applied Mathematics Seminar

 

Date and Time: 
Wednesday, June 24, 2020 - 12:00pm to Thursday, June 25, 2020 - 11:55am
Venue: 
Zoom

Statistics Department Seminar presents "A Bayesian approach to contamination removal in molecular microbial studies"

Topic: 
A Bayesian approach to contamination removal in molecular microbial studies
Abstract / Description: 

High-throughput sequencing (HTS) allows the quantification of non-culturable microbial organisms in human health and disease states, including infectious diseases. However, contaminating nucleic acids (DNA) from external sources may lead to misidentification of a taxon's provenance. Sequencing controls can help to identify most of these contaminants through the use of statistical mixture models. We propose a Bayesian reference analysis based on a hierarchical model for the observed data, that infers the true intensities of a specimen's microbial DNA in the presence of microbial DNA contamination. By using the partial information about contamination intensities available in negative controls, we define a marginal likelihood and reference prior for the true intensities. Then, we obtain a marginal posterior distribution for the true intensities.

In this talk, I will present the performance of the contamination removal method in the dilution series of the standard ZymoBIOMICS microbial community. I will also demonstrate our approach on two different low-biomass plasma specimens datasets. Our method is available as an open-source R package on Github. In addition, to identify contaminant sources, we provide a topic modeling approach to infer contaminant topics.

Date and Time: 
Tuesday, June 2, 2020 - 4:30pm
Venue: 
Zoom ID 935 0733 5349 (locks at 4:40pm PST)

Probability Seminar presents "Shuffling ancestries"

Topic: 
Shuffling ancestries
Abstract / Description: 

Ranked tree shapes and ranked genealogies are binary tree structures commonly used in biological areas. These trees are used to model the ancestral history of a sample, typically a sample of DNA or RNA sequences. We will discuss two representations of ranked tree shapes as constrained matrices of integers and ordered matchings. We exploit these representations to define an ergodic Markov chain on the space of ranked tree shapes with uniform stationary distribution. We will study its mixing time and compare to other related work.

This is based on joint work with Mackenzie Simper.

Date and Time: 
Monday, June 15, 2020 - 4:00pm to Tuesday, June 16, 2020 - 3:55pm
Venue: 
Zoom

Probability Seminar welcomes Yuriy Nemish

Topic: 
TBA
Abstract / Description: 

We consider general self-adjoint rational functions in several independent random matrices whose entries are centered and have constant variance. Under some numerically checkable conditions, we establish for these models the optimal local law, i.e., we show that the empirical spectral distribution on scales just above the eigenvalue spacing follows the global density of states which is determined by free probability theory. Moreover, in the framework of the developed theory, we study the density of transmission eigenvalues in the random matrix model for transport in quantum dots coupled to a chaotic environment.

This is a joint work with Laszlo Erdös and Torben Krüger.

Date and Time: 
Monday, June 8, 2020 - 4:00pm to Tuesday, June 9, 2020 - 3:55pm
Venue: 
Zoom

Probability Seminar presents "Optimal delocalization for generalized Wigner matrices"

Topic: 
Optimal delocalization for generalized Wigner matrices
Abstract / Description: 

We consider eigenvector statistics of large random matrices. When the matrix entries are sampled from independent Gaussian random variables, eigenvectors are uniformly distributed on the sphere and numerous properties can be computed exactly. In particular, it can be shown that extremal coordinates are no larger than $C\sqrt{\log N/N}$ with high probability.

There has been an extensive amount of work on generalizing such a result, known as delocalization, to a more general entry distribution. After giving a brief overview of the previous results going in this direction, we present an optimal delocalization result for matrices with sub-exponential entries for all eigenvectors. The proof is based on the dynamical method introduced by Erdös–Yau, the analysis of high moments as well as new level repulsion estimates which will be presented during the talk.

This is based on a joint work with P. Lopatto.

Date and Time: 
Monday, June 1, 2020 - 4:00pm to Tuesday, June 2, 2020 - 3:55pm
Venue: 
Zoom ID 959 3090 9831 (locks at 4:10pm PST)

Pages

Subscribe to RSS - Statistics and Probability Seminars