Colloquium Series

The Fall 2021 colloquia are held according to the posted schedule via Zoom and in Haag Hall Room 313. The Colloquium Series is open to public, and our graduate students are highly encouraged to attend. For more information please contact Dr. Yong Zeng Zeng, ZengY@umkc.edu

Fall 2021 

 

 

Friday November 12, 3:00–4:00 pm, Haag Hall Rm 313

The Most Mathematical Game

Speaker: Mr. Bill Shillito, Department of Mathematics, Oglethorpe University

Audience: Any student interested in mathematics. No previous knowledge is necessary.

Abstract: Lots of games involve numbers. But what if it were the other way around, and numbers could be made in terms of games? We'll explore this question through a game called Hackenbush, famously investigated by John H. Conway, as we see how it gives rise to an entire system of numbers (and pseudo-numbers) — some enormous, others minuscule, and still others fuzzy. All of these will be viewed through a lens of how one can, quite literally, play with mathematics.

Bryan Harris will introduce the speaker.

 

--------------------------------------------------------

 

Friday November 19, 3:00–4:00 pm

Join Zoom Meeting: https://umsystem.zoom.us/j/98040675707?pwd=R2s1bkJDTE0rSVNuZWRBQlJhRENiQT09

Statistical Aspects in Program Evaluation: Building Bridge between Disciplines

Speaker: Dr. Farzana Sultana, Greenbush, Southeast Kansas Education Service Center

 

This talk will showcase the application of statistical approaches in the field of Program Evaluation. Statistical methods will include parametric and non-parametric, univariate and multivariate analysis of quantitative data in statistical software, like SPSS, SAS, and R, with various data visualization and reporting software, like Power Bi and mySidewalk. This presentation will exhibit Dr. Sultana’s work from a wide range of evaluation fields, including independent consultancy, instructional method, professional development programs, educational projects, architectural design, community programs, assessment development ensuring item validity and reliability, and so on.

 

---------------------------------------

Friday December 3, 3:00–4:00 pm

A Joint Nested Random Effects Model for Metagenomic Trajectory Analysis with Disease Outcome in Matched Sets

Join Zoom Meeting: https://umsystem.zoom.us/j/98040675707?pwd=R2s1bkJDTE0rSVNuZWRBQlJhRENiQT09

Speaker: Dr. Qian Li, Biostatistics Department, St. Jude Children's Research Hospital

 

Many pediatric studies had found operational taxonomy units (OTUs) in children's gut microbiota predictive of hosts' growth and disease status. Most metagenomic trajectory analyses aimed to compare OTUs between exposure groups instead of disease outcomes and did not use matched sets. We proposed a joint mixed effect model (JointMatch) to detect OTUs longitudinally associated with disease outcome in matched participants. The disease status was predicted by logistic regression with two random effects nested in matched sets and subjects. The longitudinal relative abundance per OTU was fitted by a zero-inflated generalized mixed-effect model with the above random effects. We used a matched-set-specific marginal likelihood and a Wald statistic to test the OTU-disease association. We simulated disease outcome and temporal high-dimensional metagenomic counts in matched sets to compare JointMatch with LMM and ZIBR, showing JointMatch yielding higher detection power and lower type I error rate. An application to the longitudinal gut microbiota of TEDDY participants showed that JointMatch identified more taxa in infant-age stool samples signaling the hosts' autoimmune status. The work is joint with Kendra Vehik, Suraj Sarvode Mothi, Yi-Juan Hu, and Jeffery Krischer.

 

 

 

Colloquium Archives