Chris Armstrong

chrisarmstrong151@gmail.com / (202) 577-2124 / Washington, DC

I work as a bioinformatician for the Gregory lab at the University of Pennsylvania, developing models for bench data and creating data visualizations. My primary research interest in the lab is the N⁶-methyladenosine (m6A) modification in A. thaliana.

I also work for the Mazumder (HIVE) lab at George Washington University as the chief engineer for the FDA-funded BioCompute Object API. The portal interface for this project can be found here and the source code for this can be found on github.

Research Interests: Combinatorics, Data Visualization, Genomics and Epigenetics, Probability and Statistics

Proficiencies (5+ years): Applied Mathematics (statistics, convex optimization, linear regression, simulation); Python (class-based programming, joblib [parallelization library], jsonschema, NumPy); R (base, data.table, doParallel and forEach [parallelization library], ggplot2, plotly); SQL; Japanese (daily-level fluency)

Familiarities (1-3 years): Sophomore biology sequence (general biology, genetics, general and organic chemistry); C++; Genomics and Epigenetics; Python (D3); Real Analysis

Academics: B.S. Mechanical Engineering, M.S. Analytics (Quantitative Track)

Mini Portfolio – Visualization Portfolio / Combinatorics Preprint / ML/JSONTopos Visualizer / Code Samples

Any work without links is due to proprietary agreements.

Visualizations

A primary focus of my research is data visualization. In particular, I am interested in producing graphs and multimedia for non-technical/specialized audiences. I aim for visualizations that accomplish two things: 1) reducing the time between results and analysis, and 2) making clear which part of a hypothesis was tested and whether or not the test indicated anything of relevance to the experiment.

Nearly all of the visualizations below were created using custom scripting requiring at least 1,000 lines in the data pipeline. That is, in addition to "out-of-the-box" graphing, I am also capable of creating new graphing paradigms from libraries such as ggplot2 and plotly.

Proficiencies:

Libraries
R - base / ggplot2 / plotly

Explainer

Armstrong, Chris D. The Anagram Formula is Uniquely Maximized in the Positive Integers. 2020. TS. The University of Pennsylvania, Philadelphia. ArXiv. The University of Pennsylvania, 2020. Web. 2020. <https://arxiv.org>. ORCIDs: 0000-0002-9236-472X (Armstrong, Christopher).

Areas: Combinatorics

Summary: Armstrong shows that maximizing solutions to the anagram formula exist uniquely in the positive integers.

Contribution: All (only author).

Explainer Video

Associated Media

Publications

My background is primarily in quantitative disciplines (mechanical engineering and analytics). In a research setting, I have applied these skills to study primarily N⁶-methyladenosine data in A. thaliana. In particular, I study how the structure/function relationship of mRNA and the m6A modifiction are related. However, this modification is well known in not only plants, but also in humans. It is associated with cancer epigenetics, being involed in all stages of tumor initiation, growth, and metastasis.

The first two preprints below are related to this modification.

Proficiencies: statistics, convex optimization, linear regression, simulation

Areas: Combinatorics

Summary: Armstrong shows that maximizing solutions to the anagram formula exist uniquely in the positive integers.

Contribution: All (only author).

Preprint

Sharma, Bishwas et al. "Plant Direct." 2020. TS. The University of Pennsylvania, Philadelphia. Plant Direct. The American Society of Plant Biologists, 2020. Web. 2020. <https://onlinelibrary.wiley.com/journal/24754455>. ORCIDs: 0000-0002-9236-472X (Armstrong, Christopher).

Areas: Plant Genomics

Summary: Sharma et al. show that A. thaliana responds to abiotic cold stress conditions with highly selective m6A deposition and mRNA expression.

Contribution: Armstrong showed that m6a deposition in A. thaliana under abiotic cold stress conditions is highly specific to partitioned sets of genes responsible for cold stress response. In particular, cold-specific m6a sites are shown to be highly selective for the stop codon and downstream positions with a deposition probability of approximately half that of background.

Preprint (My Contribution)

Armstrong, Chris and Matthew Hieseger. "UDC Internal Document." 2014. TS. The University of The District of Columbia. UDC Internal Document. The University of the District of Columbia, None. Print. 2020. None. ORCIDs: 0000-0002-9236-472X (Armstrong, Christopher).

Areas: Applied Statistics, Adult Education

Summary: Armstrong and Hieseger conducted a review of mathematics education at an adult workforce center.

Contribution: Armstrong mentored a statistician intern (who went on to become an actuary) while he and the intern conducted a review of the mathematics curricula in a vocational setting. Together they performed a statistical analysis to determine the mathematical aptitude of incoming students, and also conducted workshops to improve mathematical literacy. Recommendations for the University were provided in the form of a report.

Report

Data

In addition to exploratory research, I am also interested in application development. I have it as a main goal to take the complex scripts that I write and make them usable for other researchers (especially those without a strong programming background). "Benchworks", listed below, is an R library that is in the process of being standardized so that it can be published on the CRAN server. The package focuses on parallelized, command line-executable analysis for bench researchers.

Proficiencies:

Libraries
Python - JSON / NumPy / Pandas / plotly
R - base / dplyr / ggplot2 / plotly / Shiny

UPenn

Benchworks

Interface

Back End

I created a data visualization app to facilitate the graphing of large and varied data sets. The app consists of a back end batch processor for data files, a metric file system, and a file coordinator. The front end consists of a Shiny interface which allows the user to customize graphing options for the batch grapher. This app makes exentive use of Shiny, jQuery, and third party libraries like shinyjs.

5000+ lines

GWU

BCO API | Link >

Interface

Back End

As the chief engineer for the FDA-funded BioCompute Object API (BCO API), I have produced a highly modular, abstracted python package. This package provides a BCO API server which can be configured with simple configuration files, and allows for several BCO API-related functions to be utilized by lab members.

2500+ lines

Georgia Tech

Data Visualizer | Link >

Interface

Back End

As the sole app developer on my team, I modified code from two third-party projects to create a chloropleth map of the United States. The user interface was designed with highly customizable search parameters so that the main data set could be subsetted to a high precision. This project made extensive use of JSON, topoJSON, and D3 capabilities, and was my first chloropleth project.

1000+ lines

Code Samples

You can view or download my code samples here.

Resumé

You can view or download my resumé here.