Subgroups of Diabetes

Precision Medicine in Type 2 Diabetes

There are two broad categories of projects under this theme:

  1. Defining Transportable Endotypes of Type 2 Diabetes
  2. Global Inequities in Precision Medicine Research for Diabetes

1. Defining Transportable Endotypes of Type 2 Diabetes

A novel approach for precision medicine for diabetes involves classifying newly diagnosed type 2 diabetes patients into data-driven endotypes based on their clinical characteristics (e.g., age at diagnosis, body mass index, HbA1c, HOMA2-IR, HOMA2-B). These endotypes vary in their genetic profile, presentation, responses to antihyperglycemic medication and risk of diabetes-related complications. However, reasons such as unavailability of HOMA2 indices in routinely collected clinical data, scarcity of data on endotypes derived from non-European populations, non-replicable cut-points for clinical variables and hard cluster assignment ignoring multifactorial nature of diabetes, preclude the widespread translation of this phenotyping approach from cohort studies to clinical practice.

This is ongoing work with Dr. KM Venkat Narayan, Dr. Joyce C. Ho and Zhongyu Li.

Challenges in defining endotypes for type 2 diabetes.


  1. Varghese 2023 Lancet Diabetes & Endocrinology Correspondence on Replicability of Endotypes
  2. Varghese 2023 Primary Care Diabetes Brief Report on Ethnic Differences of Endotypes
  3. Varghese 2021 Diabetes Technology & Therapeutics on CGM-derived Endotypes

Getting Involved

These projects are ideal for advanced masters or doctoral students who want hands-on experience with cohort studies and electronic health record data.

If you are interested in adult-onset type 2 diabetes:

  1. Classify endotypes of newly diagnosed adult type 2 diabetes from cohort studies using variables available in electronic health records
  2. Visualize distribution of endotypes in EHR systems by county, race-ethnicity, sex, and age groups
  3. Describe trends in prescribed medications, clinical parameters, and risk of complications among EHR-derived newly diagnosed type 2 diabetes endotypes.
  4. Develop a ‘2-year risk’ multi-class prediction model for phenotype membership using electronic health records from two clinical research networks and evaluate domain generalization

If you are interested in youth-onset type 2 diabetes:

  1. Develop a novel classification of newly diagnosed, youth-onset type 2 diabetes using clinical characteristics from cohort studies, and compare the classification with adult-onset endotypes
  2. Describe the longitudinal association of T2DM endotypes among youth-onset type 2 diabetes and Michigan Neuropathy Screening Instrument (MNSI) scores
  3. Describe the association of youth-onset T2DM endotypes and distal symmetric polyneuropathy phenotypes based on the Michigan Neuropathy Screening Instrument (MNSI)


  1. Proficiency in R or Python
  2. Required Coursework: EPI 560 Epidemiologic Methods IV, EPI 568 Bias Analysis, Longitudinal Analysis (BIOS 502 or BIOS 525 or BIOS 526)
  3. Ideal Coursework: CS 534 Machine Learning

2. Global Inequities in Precision Medicine Research for Diabetes

While non-Europid populations contribute to nearly 80% of the global diabetes burdens, they are grossly underrepresented in precision medicine research. We undertook an electronic search of literature in PubMed from 2010 to 2023 to identify studies in precision medicine in diabetes in populations from East Asia (EA), Latin America & Caribbean (LAC), Middle East & North Africa (MENA), South Asia (SA) and South East Asia & Pacific Islands (SEAP). We additionally included Central Asia, Central & Eastern Europe, and Sub-Saharan Africa in the subsequent analysis.

This is ongoing work led by Aamna Soniwala and Sophia Kim.

Funding: Emory Global Diabetes Research Center

Summary of findings presented at American Diabetes Association's Scientific Sessions 2023.

Getting Involved

These projects are ideal for advanced masters or doctoral students, who are independently skilled, to apply natural language processing techniques.

  1. Identifying key themes using NLP


  1. Soniwala 2023 Diabetes


  1. Proficiency in Python and Natural Language Processing