Harvard

Research

Introduction

After I finished my PhD, I did postdoctoral research at the Harvard School of Public Health in Boston.

I researched the molecular biology of nutrition and metabolism during this time. I did some work to help illuminate the links between food molecules and physiology. I won grant funding and collaborated with researchers in academia and pharma.

Nutritional screening

Overview

Chronic diseases with established nutritional risk components, including obesity, diabetes, heart disease, stroke, and cancer, are responsible for millions of deaths each year and constitute an enormous global health burden¹. While diseases are strongly connected to food choices, the human diet remains poorly characterized.

Past study of the molecular components of foods led to the discovery of the essential nutrients, and has had significant positive impact on public health. In addition to essential nutrients like vitamins and minerals, foods, plants, and dietary supplements contain thousands of biologically active compounds ("dietary bioactives") that act at the cellular level to impact health. The overall objective of this work was to develop novel tools to identify the cellular effects of dietary bioactives, with the long-term goal of harnessing the power of dietary bioactives to improve human health.

Past advances in molecular nutrition had a positive impact on public health

A connection between diet and health has been postulated for thousands of years. In Greece during the fifth century BCE, Hippocrates recommended that "food be thy medicine," and hypothesized salutary properties of foods such as garlic². Seaweed was first used as a folk remedy for goiter (enlargement and impaired function of the thyroid gland), but the reason for its effectiveness was not known.

Many years later, advances in medicine and physiology led to renewed interest in the physiological effects of foods. In his famous 1746 trial, British naval surgeon James Lind successfully treated scurvy in sailors with citrus fruits³. Thomas Percival wrote on the ability of cod liver oil to treat the bone condition known as rickets in 1789⁴. While these early discoveries were important advances for nutrition and medicine, knowledge remained anecdotal, fragmented and incomplete, and the scientific tools available to researchers at the time did not allow them to derive causal insight from their experiments. Lind incorrectly hypothesized that sea air caused scurvy⁵.

Advances in the field of biochemistry allowed scientists to study foods at a molecular level of detail, leading to rapid discovery of the essential nutrients. The antiscorbutic factor in citrus fruits was isolated, characterized, and named vitamin C. Vitamin D was identified as the necessary factor for prevention of rickets. The chemist Bernard Courtois isolated iodine from seaweed, enabling the physician David Marine to conduct clinical trials showing that iodine was the essential molecule needed to treat goiter⁶. Numerous other discoveries were made in this time period, with at least fifteen Nobel Prizes awarded for the study of essential nutrients⁷ (also see table). Biochemistry gave scientists the tools to advance nutrition research, and this knowledge had a major impact on public health through prevention of deficiency diseases. It was only after we began to understand the biochemical composition of foods that we could determine dietary nutrient requirements to prevent deficiencies and promote optimal physiological function.

New methods are needed to identify the molecular targets of nutrients and dietary bioactives

The American Society for Nutrition has identified several priority areas, including an enhanced understanding of food composition, incorporation of this information into more comprehensive nutrition databases, and use of high-throughput omics techniques⁸. New methods are needed in order to make progress on these priority areas. It is not enough to rely on borrowed techniques from molecular biology and analytical chemistry.

As one example, vitamin D status is determined by measuring circulating 25-hydroxy vitamin D, which is inadequate for three reasons. First, measurement methods are inconsistent⁹. Second, the circulation is not the major site of action of vitamin D. Third, 25-hydroxy vitamin D is not the functional form of the vitamin. Measurements of nutritional status are thus several steps removed from a functional mechanism.

And foods are more than just essential nutrients. In addition to essential nutrients like vitamins and minerals, foods and dietary supplements contain thousands of biologically active compounds ("dietary bioactives") that act at the cellular level to impact health. Even a seemingly simple cup of coffee can contain a thousand such dietary bioactives¹⁰. Evidence for the health impact of dietary bioactives has accumulated to the point that they can now be incorporated into dietary guidelines. The first dietary bioactive guideline was published for flavan-3-ols (flavonoids)¹¹.

Aims

Lipids are a class of compounds that includes fatty acids and bile acids. The endoplasmic reticulum (ER) is a cellular organelle that is particularly important for the metabolism of lipids and is involved in the pathogenesis of chronic metabolic diseases. While lipid synthesis at the ER has been widely studied, little is known about how the ER handles the many lipids from the diet, and how the response of the ER to dietary lipids impacts health.

The aims were to:

1. Express ER reporters in new cell types. In order to detect the cellular effects of lipids, this work focused on indicators of ER function. We developed reporters that enabled real-time monitoring of ER function in live cells and aimed to introduce these reporters into liver, adipose, pancreatic, and immune cells.
2. Create and utilize a novel cellular screening system to detect the effects of lipids on the ER. We aimed to create a microfluidic device to improve on standard cell culture screens. This device, approximately the size of a microscope slide, was envisioned with chambers for cells expressing the ER reporters and channels to flow culture medium and treatments to the cells. We aimed to use this microfluidic device for high-throughput screening of lipids and continuous sampling of the cell culture medium for evaluation of the dynamic response of the ER.

Conferences and collaborations

Conferences are helpful for finding research collaborators. I pursued some collaborations in order to advance the second aim of creating a novel cellular screening system. Through a meeting at the SelectBio Organ-on-a-Chip World Congress in 2016, I established one collaboration with an MIT lab that was working on digital microfluidics¹², also called electrowetting¹³. This technology uses electricity to move liquid droplets around. Digital microfluidic systems have some notable advantages over other microfluidic approaches. One advantage is reusability. Unlike microfluidic systems made of disposable PDMS (polydimethylsiloxane), the device can be reused an unlimited number of times. Another advantage is directional independence. Unlike microfluidic systems that flow liquids in one direction, digital microfluidic devices can be set up to move droplets in any direction. This directional independence means that we could potentially move cells around on the device to study physiological processes. For example, a colony of adipocytes (fat cells) could be cultured in droplets, and the device could induce influx and efflux of immune cells like macrophages to and from the adipocytes. Interactions between fat cells and immune cells are important in the development of obesity, diabetes, and other metabolic disorders¹⁴.

We did some preliminary experiments with the digital microfluidic device and had fun working on it together. The collaboration paid off later in my career as well -- through this collaboration, I met another friend who ended up getting me one of my tech jobs.

I also presented some preliminary plans for the nutritional screening project at the Experimental Biology (EB) 2017 conference in Chicago. You can see my conference poster above.

Lab management

The lab was frequently in disarray and needed a competent lab manager. I volunteered to serve as interim lab manager and served in this capacity for about eight months. I worked with the other lab members to achieve numerous improvements to lab infrastructure and culture. I also encountered a set of lab management pain points that could be addressed with better tech.

Inventory management

• Led a clean-out and decontamination of the lab cold room (walk-in refrigerator): The cold room had not been properly maintained and was full of mold and expired products. I worked with lab members to review inventory, discard or disinfect materials, and move materials to temporary storage. I had the room professionally decontaminated. I then moved materials back in, and established guidelines to keep the cold room cleaner.
• Consolidated and simplified the distribution of common supplies in the lab
• Created and curated a database of essential lab supplies and kept them stocked
• Coordinated shipment of biological materials around the world
• Prepared a shipping protocol to assist others in the department

Procurement

• Performed ordering, receiving and organization of all lab supplies
• Saved thousands of dollars with strategic procurement
• Established a budget tracking system for lab members
• Coordinated educational seminars and product demos

Safety

• Oversaw chemical, radioactive, and biological safety, including adherence to regulatory guidelines and disposal of waste
• Successfully passed a chemical safety inspection and addressed its recommendations
• Advocated for improved lab climate control: Harvard had been turning off the air (both circulation and cooling) to our lab outside of business hours, resulting in oppressively hot and stuffy conditions for the lab members. I brought together administrators to change the climate control policy.

Transition into tech

During my postdoc, I began to see that the science career system does not incentivize reproducibility. In science, reproducibility occurs when different scientists do the same experiment and get results that agree. The scientific community is experiencing what has been called a "reproducibility crisis" or "replication crisis," in which published research can't be reproduced by multiple labs, or oftentimes even repeated within the same lab by the same person. Cell Reports published an instructive anecdote on the topic¹⁵, and the book "Rigor Mortis" by Richard Harris¹⁶ provides a useful summary of the crisis.

I started thinking about how software, lab automation, and other technology tools could help with reproducibility. Effective lab data capture certainly seems important for reproducibility. Molecular biologists perform complicated experimental protocols with many steps, and each step can affect the outcome. It's difficult to document experiments in enough detail so that other scientists can learn from, and reproduce, our work. I could see the importance of reproducible protocols in other fields as well. For example, the surgeon Atul Gawande wrote a book called "The Checklist Manifesto"¹⁷ describing their research¹⁸ about surgical checklists, a form of protocol or documentation. Distributing a checklist to surgical team members reduced patient deaths by half. Couldn't software help us document and reproduce our work?

I started using an electronic lab notebook (ELN) to replace the antiquated paper lab notebooks my lab at Harvard was still using. I used an iPad to capture data in the ELN during my experiments. It was better than paper, but it wasn't a smooth transition. On a Saturday afternoon on which our ELN was experiencing an extended cloud outage, I had an epiphany. I realized software skills would empower me to alleviate problems like this and improve productivity for other scientists in the future. I decided to learn software engineering and start a career in tech.

---

Footnotes

1. CDC. National Vital Statistics Reports. ↩

2. Rahman K. Historical perspective on garlic and cardiovascular disease. Journal of Nutrition 2001. ↩

3. Manchester KL. An orange a day keeps the scurvy away. Trends in Pharmacological Sciences 1998. ↩

4. Jones G. Chapter 18: Vitamin D (pp.278-292). In: Modern Nutrition in Health and Disease (11th edition). Lippincott Williams & Wilkins 2012. ↩

5. Carpenter KJ. A short history of nutritional science: part 1 (1785-1885). Journal of Nutrition 2003. ↩

6. Carpenter KJ. David Marine and the problem of goiter. Journal of Nutrition 2005. ↩

7. Carpenter KJ. The Nobel Prize and the discovery of vitamins. NobelPrize.org 2004. ↩ ↩²

8. Ohlhorst SD et al. Nutrition research to affect food and a healthy life span. Journal of Nutrition 2013. ↩

9. Sempos CT et al. Vitamin D status as an international issue: national surveys and the problem of standardization. Scandinavian Journal of Clinical and Laboratory Investigation 2012. ↩

10. de Mejia EG, Ramirez-Mares MV. Impact of caffeine and coffee on our health. Trends in Endocrinology & Metabolism 2014. ↩

11. Crowe-White KM et al. Flavan-3-ols and cardiometabolic health: first ever dietary bioactive guideline. Advances in Nutrition 2022. ↩

12. Choi K et al. Digital Microfluidics. Annual Review of Analytical Chemistry 2012. ↩

13. Wheeler AR. Putting electrowetting to work. Science 2008. ↩

14. Rosen ED, Spiegelman BM. What we talk about when we talk about fat. Cell 2014. ↩

15. Hines WC et al. Sorting out the FACS: A devil in the details. Cell Reports 2014. ↩

16. Harris R. Rigor Mortis (1st edition). Basic Books 2017. ↩

17. Gawande A. The Checklist Manifesto (1st edition). Metropolitan Books 2009. ↩

18. Haynes AB et al. A surgical safety checklist to reduce morbidity and mortality in a global population. New England Journal of Medicine 2009. ↩