-
Notifications
You must be signed in to change notification settings - Fork 2
/
research.html
39 lines (36 loc) · 3.17 KB
/
research.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
---
layout: default
title: 'Research'
---
<h1 style="margin-top:0;">Research</h1>
<p class = "lead"><small>My research objectives emphasize the development of user-friendly, statistically sound, and open-source software that addresses practical scientific challenges and empowers non-statisticians/data scientists to work more directly with their data.
I deeply value interdisciplinary collaboration, and welcome you to <a class = "active" href="contact.html">contact me</a> if you are interested.
You can find detailed information about my publications, software, and other work on <a class="active" href="cv.html">my CV</a>.</small></p>
<h4 style="margin-top:0">Statistical Modeling of the Microbiome</h4>
<p class="content" style="padding:0px;margin:0px 20px 0px;">
I work with <a href = "https://faculty.washington.edu/dwitten/">Daniela Witten</a> and
<a href = "https://faculty.washington.edu/adwillis/">Amy Willis</a> to develop methods for proper statistical analysis of the microbiome.
I am the developer of corncob, a regression model and hypothesis testing procedure developed
to address the challenges of microbiome data, implemented as an R software package.
Our proposal uses observed absolute abundance information to
make inference about the latent microbial population.
The method is published in <i>Annals of Applied Statistics</i>
(available <a href = "https://projecteuclid.org/euclid.aoas/1587002666">here</a>) and the software is available
<a href = "https://github.com/bryandmartin/corncob">here</a>.
<br/><br/>
I am currently working on investigating the impact of efficiencies, or systematic differences between microbial population relative abundances and the expected observed sample relative abundances (e.g. from preferential sampling, PCR amplification, etc).
We are investigating the impact of model misspecification from efficiencies on differential abundance methods that do not account for them.
We are also exploring models that are able to able to conduct valid inference on parameters of interest for differential abundance even in the presence of efficiences.
<br/><br/>
I also helped develop DivNet,
a method for estimating alpha and beta diversity of microbial communities when taxa interact via an ecological network.
The method is published in <i>Biostatistics</i> (available <a href = "https://academic.oup.com/biostatistics/advance-article/doi/10.1093/biostatistics/kxaa015/5841114">here</a>) and
the software is available <a href = "https://github.com/adw96/DivNet">here</a>.
</p>
<br/>
<h4>Small-Area Spatio-temporal Modeling</h4>
<p class="content" style="padding:0px;margin:0px 20px 0px;">
I also work on developing spatio-temporal models for under-five mortality estimation with complex survey data
with a group led by <a href = "https://faculty.washington.edu/jonno/">Jon Wakefield</a> and
<a href = "https://sociology.osu.edu/people/clark.2962">Sam Clark</a>. The method is published in PLoS One (available <a href = "https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0210645">here</a>) and our software, SUMMER, is available on CRAN <a href = "https://CRAN.R-project.org/package=SUMMER">here</a>.
</p>