Melinda Higgins, Ph.D.

About

I’m a Biostatistician and Chemometrician - both of which represent my passion and focus applying statistics, multivariate pattern recognition, modeling and programming for wide and diverse research areas. I love learning new analysis methods and approaches and writing the code to make it happen. My latest endeavors include using reproducible research and dynamic document procedures to make my data analyses streamlined, easy to execute and well annotated.

Current [2007 - ]

I am currently a senior bio-statistician and associate research professor in the Office of Nursing Research at Emory University in the Nell Hodgson Woodruff School of Nursing. I primarily collaborate with our Nursing researchers on their clinical trials, but also support research in the School of Medicine and at Yerkes National Primate Research Center.

I most often use multilevel mixed linear and generalized linear models for longitudinal data analyses - usually comparing a treatment to usual care groups focusing on “group-by-time” interaction effects. However, I have also recently employed instrument response theory and factor analysis exploring latent variable structure for disease knowledge tests and symptom assessment. I have also begun exploring network graph theory and visualization techniques for visualizing multivariate associations and correlation structure among social, psychological, behavioral measures as well as bio-markers and “-omics” outcomes.

My Background, Experience and Interests

[1975-1985] My early years - Science and Math

Since my earliest years in grade school and high school I have always been passionate about science and math. In high school I was mentored by an excellent Chemistry teacher who was a PhD research chemist with expertise in organic chemistry. It was in high school that I first worked with scientists at the Fernbank Science Center and at Georgia Tech. They helped me with my science fair projects where I first used gas chromatography and did my first initial data analyses for modeling chemical composition. I also began learning about computers and programming. I learned how to work with a TRS-80 (aka “trash-80”) and began running computer programs in BASIC via punch cards using a handset modem and a card reader.

[1985-1989] College - Dual Major Math & Chemistry

After high school, I attended Saint Andrews Presbyterian College (now Saint Andrews University) in Laurinburg, NC from 1985-1989. While in college I double majored in math and chemistry. My math classes covered Calculus, differential equations, and statistics. In chemistry, I took many classes covering analytical chemistry, organic chemistry, natural product synthesis, physical chemistry, quantum mechanics and physics. I was also fortunate to work hands-on with many instruments including: gas and liquid chromatography, NMR (nuclear magnetic resonance), IR spectroscopy, UV spectroscopy, mass spectrometry, bomb calorimetry, pH meters, melting point (Mel-Temp) as well as glassware configurations for distillation and other chemical synthesis. I also learned how to program in PROLOG logic programming language, PASCAL and some FORTRAN with brief introduction to OOP (object oriented programming). I worked on mainframe computers running UNIX and on PC’s running DOS and learned batch processing and coding.

[1985-1989] Student Apprentice at Georgia Tech Research Institute (GTRI) - Indoor Air and Environmental Exposure Lab

While home in Atlanta during summers and holiday breaks during my college years from 1985-1989, I worked as a student employee at the Georgia Tech Research Institute. I worked in an environmental exposure lab run by Dr. Charlene Bayer, current Chief Scientist at Hygieia Sciences LLC, and Dr. Marilyn Black (who left GTRI and founded Air Quality Sciences and the GREENGUARD Environmental Institute). I analyzed samples of indoor air, water, soil and commercial products. I analyzed these various media for many types of contaminants. However, primary analytes of focus included pesticides, asbestos, second-hand smoke and volatile organic compounds (VOCs). The lab was well known for their work on “sick building syndrome” and modeling indoor air contamination using real-time simulation and data collection with custom built room sized air exchange chambers.

I gained a lot of experience working with gas chromatographs using a variety of detectors including FID (flame ionization detectors), NPD (nitrogen phosphorous detector) and ECD (electron capture detector). I also worked with hybrid GC/MS (gas chromatography/mass spectrometry) which produced large files with hundreds to thousands of data points for the complicated organic chemical profiles from the samples. I also used liquid and ion chromatography as well as stand alone mass spectrometers (SCIEX) with various modes of input (like Ion Spray) and detectors. Interestingly enough, these mass specs were linked to Macs, which gave me the opportunity to work with the Mac operating system and interface as well.

Needless to say working with these many instruments and types of analyses and a wide range of analytes, I became increasingly interested in data manipulation and analyses - especially on how to best handle and approach these multidimensional datasets. By the time I completed my undergraduate degree (Bachelor’s of Science with dual-major in math and chemistry), I realized that I wanted to continue my education and get my PhD for several reasons - but mainly because I was ready to do more than prepare and run analyses and I wanted to learn more data analysis and move away from wet chemistry and do more using computers.

[1989-1994] Graduate School (part 1) - Chemometrics, Experimental Design and Optimization

Given my work experiences at GTRI, I found the field of Chemometrics fit the bill for what I wanted to do. Up to this point given all of my work with numerous instruments for analyzing chemical samples, I was well trained to pursue a degree in Analytical Chemistry. Analytical Chemistry is the discipline of finding optimal ways of breaking down samples into their components to figure out what is there and how much. This can be accomplished chemically by working on separation science (more wet chemistry) OR numerically (yea!) through multivariate pattern recognition and classification approaches. This is Chemometrics. In 1989 there were not many researchers let alone schools/universities with a focus in Chemometrics. One of the most well known Chemometricians in the US was Dr. Bruce Kowalski at the University of Washington in Seattle, WA. Dr. Kowalski’s research group was very large with multiple post docs and numerous graduate students - where I worried I’d disappear into the morass of the big group - plus it was over 2500 miles away from home. I also found Dr. Stephen Morgan at the University of South Carolina who was also well known for his work and short courses in Experimental Design and Optimization methods as well as work in Chemometrics. So to USC I went.

When I began my PhD. work, my initial research focused on using deuterium to selectively tag stereo-isomers of the hexose sugars (alditol hexaacetates) for optimal separation of these sugars using mass spectrometry with numerical classification and separation via factor analysis methods. At the time, we were still running DOS based computers - so to accomplish the factor analyses, I used the LAPACK linear algebra package routines. I wrote all of the codes in PASCAL to accomplish the analyses. Then we slowly upgraded computers - first to Window-DOS then eventually to Windows 3.x. After getting Windows running we also had access to the MATLAB software and programming language which made matrix algebra routines a breeze. You could accomplish singular value decomposition in 1 line of code!! I was converted. I abandoned PASCAL and moved all of my PhD research into MATLAB.

What I was missing, however, was a solid foundation in matrix algebra and statistical theory. So, while pursuing my PhD in Chemometrics, I also began taking as many math and stats courses as I could to support my research. By the time I finished my PhD. dissertation in 1994, I was within 9 course hours of having a Masters in Statistics. So, after I graduated with my PhD I immediately enrolled in the Statistics department to finish my Masters in Statistics.

[1990-1995] Graduate School (part 2) - Statistics

While I had been taking math and stats courses since the early 90’s, I officially enrolled in the Statistics Department at USC in 1994. I choose Dr. Don Edwards as my advisor and began my Master’s Thesis. The focal point of my thesis was on applying novel experimental design techniques for optimizing measurements related to estimating enzyme kinetics coefficients. My approach used the Michaelis-Menten equation. I worked with researchers in the Biology department to collect data and run experiments.

In my math and stats courses, I learned SAS and how to run it on a mainframe using JCL batch code as well as later how to run SAS in UNIX and on PCs. I also learned S-plus running in UNIX. The majority of my thesis was accomplished using S-plus. It was later while working at GTRI that I switched to R from S-plus, but I already had a solid foundation.

[1995-2011] Employment at Georgia Tech Research Institute (GTRI)

After completing graduate school in 1995, I began working at GTRI as a technical consultant for Dr. Ted Doll on his GTV (Georgia Tech Vision) model, which was a general, high fidelity model of visual performance based on the principles of human visual processing. GTV was incorporated into a signature analysis tool called Visual and Electro-Optical (VISEO) detection for the Army Aviation and Troop Command, Applied Technology Directorate (ATCOM/AATD). My contribution was writing the Discriminant Analysis codes for target classification.

While working for Dr. Ted Doll, Dr. Charlene Bayer hired me as a full-time employee working in the Environmental Analysis branch of the Electro-Optics, Environment and Materials Laboratory at GTRI. I picked up my previous work when I was a student working with Dr. Bayer doing indoor air analyses and related work. However, after a few years I was ready to expand my research into other areas of spectroscopy, pattern recognition and modeling.

In about 1997, I moved into the Electro-optics division of EOEML. I moved over to work with Dr. Bill Owens, Dr. Bob Hyde and Dr. Nick Faust. I got a security clearance and worked on a number of programs exploring target detection and classification as well as modeling aircraft and missle “signatures” looking at spectroscopic patterns and optimizing filters associated with passive signatures from reflective materials and active signiatures from heat exhaust and other sources. During this time I worked mostly in UNIX and learned many useful programs and software such as: awk/gawk; sed; PERL; vi; emacs; etc. I also programmed almost exclusively in FORTRAN as this was the main language that most of the DOD (Department of Defense) signature modeling routines were written.

I also began to work more with Dr. Nick Faust.

Contact

Melinda Higgins, PhD
Associate Professor Research
Senior Bio-statistician
Emory University
School of Nursing - Office of Research
melinda.higgins@emory.edu
404-727-5180

Website Colophon

I designed and developed this site using:

This site is hosted by github pages and version controlled via Github for Windows client.

Original Site Source available by Gaston Sanchez at github and you can fork my site at Melinda Higgins at github.

Unless specified otherwise, content on this site is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

REFs

  1. Doll, T.H., McWhorter, S.W., Schmieder, D.E., and Wasilewski, A.A. (1995). Simulation of Selective Attention and Training Effects in Visual Search and Detection, in Vision Models for Target Detection and Recognition, Peli, E. editor. World Scientific Publishing Co. River Edge, NJ.

  2. Doll, T.H., McWhorter, S.W., Schmieder, D.E., Hertzler, M.C., Stewart, J.M., Wasilewski, A.A., Owens, W.R., Sheffer, A.D.Galloway, G.L. and Herbert, S.D. (1997). Biologically-based vision simulation for target-background discrimination and camouflage/lo design. Paper No. 3062-29 in Targets and Backgrounds: Proceedings of the International Society of Photo-optical Instrumentation Engineers, Watkins, W.R. and Clemen

Random bits

Blog

My initial blog was located at Github Blog

About this website

This website is …………

colophon

SAS Univ    Web1 github    Twitter    ResearcherID    GoogleScholar   

test emojis

Test a few random emojis :heart_eyes: :sparkles: :octocat: :large_blue_circle: :blush: and a unicode 🌅