Co/Wy Chapter - Winter Meeting

When:  Feb 20, 2016 from 15:00 to 19:00 (MT)
Associated with  Colorado-Wyoming Chapter

Details for the Colorado/ Wyoming Chapter of the American Statistical Association Meeting are provided below.  The event is being presented as an alternative to our annual spring and fall meetings which occurred during the work week.  There will be a cash bar and hors d'oeuvres – courtesy of the CSU Statistics Department.  The event is free and open to all, but please RSVP to Katy Koehler (koehler@stat.colostate.edu) by 2/15 so that an appropriate amount of food can be ordered.
 
General Schedule.

3 pm  - First Talk – An Overview of Microbiome Data Analysis.

4 pm – Second talk - Statistical Models with Linear Inequality Constraints.

5 pm – Hors d'oeuvres and cash bar

7 pm.  If folks are interested in more food or drink we will chose an off-campus location to reconvene.


Location: Grey Rock room, Lory Student Center

Parking: it’s free on Saturday in most A lots.  Here is a CSU Campus Map.  We recommend people park on lot 310, in the NE quadrant of the map, and the closest to the Lory Student Center

Inclement Weather

If the event is cancelled because of inclement weather, we will do so Friday afternoon, February 21.  We will send out emails and post the cancelation on our website.


Abstracts

An overview of microbiome data analysis.

S. Huzurbazar (with E. Smirnova, E. Jackson, A. Menon, K. Williamson)
Department of Statistics, University of Wyoming

Microbiome data obtained from sequencing the 16S rRNA gene consist of a large number of variables (taxa) and usually a small number of samples, resulting in “big p, small n” data. Ordination methods based on eigen-decompositions of the data matrix are commonly used to achieve dimension reduction. These methods can be expressed using a general framework by applying an appropriate matrix transformation to the raw data followed by eigen-decomposition of the transformed matrix.

In the process of working with the NIH Human Microbiome II project at Virginia Commonwealth University, we had to become familiar with ordination methods based on the duality diagram approach (with origins in ecology), with the R package ade4, and to contend with the effects of common practices such as filtering and normalization of the raw data. In the process, we produced a survey of the duality diagram approach in terms that are more familiar to statisticians, a vignette of ade4 with examples relevant for human microbiome data, some novel perspectives on interactively visualizing more than 3 components of eigen-decompositions using R and the QIIME pipeline, and the effects of normalization and bias adjustment of ordination based analysis. I will present an overview of human microbiome data and a summary of our work.


Statistical Models with Linear Inequality Constraints


Mary Meyer, Professor, Colorado State University

 

Abstract:  In statistical models involving parametric or non-parametric function estimation, accuracy and precision can often be enhanced by incorporating a priori information in the form of inequality constraints.  For example, if a growth curve must be increasing and concave, or the probability of remission is known to be decreasing in tumor size, or the treatment effects can be assumed to be at least as large as the placebo effect, these assumptions can be formulated as constraints on the parameters or the spline coefficients.   Some background in estimation and inference with these constraints will be given, as well as some practical applications and examples using R packages.


Spring Meeting - Friday April 15th, 2016 , NCAR - Boulder  A call for speakers.

The Colorado Wyoming Chapter of the American Statistical Association's Spring Meeting  will be at the National Center for Atmospheric Research in Boulder on Friday April 15th. 

This is an invitation for speakers.  While everyone is invited to speak, we would like especially encourage students from our colleges and universities to take this opportunity to speak to a supportive crowd.  Following the format from last year, all talks will be ~15 minutes.    If you are interested in speaking - please send a brief abstract to Matt Pocernich at (Matt_Pocernich@hotmail.com ) by February 15th. This is an excellent opportunity to give share a talk with a friendly audience.  Don't be shy.

Details regarding the rest of the meeting will be made available in mid-March.  Generally speaking, the meeting will run from 9:30 until ~4, with ample time for lunch.  We will arrange a tour of the VisLab or super computer room as an optional lunch activity and likely some of us will go to the Southern Sun for a beer afterwards.  If you haven't visited the NCAR's Mesa Lab, it is possibly the nicest place on the front range to spend a Friday in April - and of course the talks will be great.

 


Location

Lory Student Center - CSU

Contact

Matthew Pocernich
(720) 936-2048
matt_pocernich@hotmail.com