Xin Fang, PhD '20
Computational Biologist, 23andMeXin Fang, PhD '20
Embedded Files
Bernhard Palsson, PhD
Distinguished Professor of BioengineeringBernhard Palsson, PhD
Larry Smarr, PhD
Professor of Computer ScienceLarry Smarr, PhD
Ecosystem Dynamics Reveal Possible Microbial Culprits for Inflammatory Bowel Disease
Ecosystem Dynamics Reveal Possible Microbial Culprits for Inflammatory Bowel Disease
Combining ecosystem dynamics, the microbiome, metabolomics, and time series analysis make a detailed case against specific strains of E. coli as having means, opportunity, and motive.
Combining ecosystem dynamics, the microbiome, metabolomics, and time series analysis make a detailed case against specific strains of E. coli as having means, opportunity, and motive.
By Asst. Prof. Benjamin Smarr, UCSD Bioengineering and Halicioglu Data Science Institute
With Xin Fang, Ph.D., 23andMe; Prof. Bernhard Palsson, UCSD Bioengineering; Prof. Larry Smarr, UCSD Computer Science.
Ecosystems that exist inside us might be easier to investigate in the lab than in the wild
Ecosystems are hard to study. They have complex dynamics, and are difficult to isolate and replicate in the lab. But ecosystems aren’t limited to macro-systems like forests or coral reefs; ecosystems hide inside our bodies too. Xin Fang received her PhD from UCSD Bioengineering in 2020 and was the lead author on three papers that put ecology into a clinical lab setting.
Xin worked with national and international collaborators, as well as across the UCSD campus, in efforts orchestrated by Prof. Bernhard Palsson. Prof. Palsson is director of UC San Diego Systems Biology Research Group, known for their approach of using large-scale computational modeling to bridge understanding from through phenotypes. Xin and her colleagues found macro-ecosystem succession dynamics among E. coli strains in the human gut microbiome, where one dominant species is suddenly deposed and replaced by another species that becomes the new normal. By extracting both metagenomics and metabolomics from a single person’s fecal time series, she was able to identify the specific strains of a single bacterial species, E. coli, undergoing succession. Xin then cross-referenced these strains’ genetics to the computationally-derived metabolomic pathway maps and isolated the mechanisms allowing the most virulent strains to succeed.
From ecosystem to petri dish with deep-omics
Going beyond normal ecosystem observation, Xin was able to reproduce these mechanisms in a petri dish to confirm their effect. Finally, Xin and the team made a translational leap, applying their ecological observations to over a hundred patients suffering inflammatory bowel disease (IBD). Among these, those that required surgery ended up with lower diversity and higher rates of succession events, suggesting that even years after operations, healthy dynamics had not been restored, and suggesting more clinical research is needed to assess the timing and efficacy of surgical interventions for IBD sufferers.
The team was able to isolate the specific strains associated with flares, E. coli B2. Using genome-scale metabolic models (GEMs), she was then able to map these strains’ unique capabilities by their unique metabolic trails, which lead her to a smoking gun: the most virulent E. coli B2 have a unique ability to digest the sugar tagatose that is found in the mucus in the human colon. As described in her BMC Systems Biology paper, this ability letsE. coliB2 digest its way to the epithelial layer, thus playing a role in triggering an inflammatory response that leads to an IBD flare.
Looking in the unique time series of samples generated by one of her faculty mentors – Prof. Larry Smarr, himself an IBD patient – Xin found that IBD flares were associated with spikes in B2 strains exactly as predicted. Even more interesting was that within the virulent strains, one specific strain would dominate at a time – a sort of Game of Thrones of E. coli. Such dynamics are common in macro-ecology, but had remained invisible without the combination of a single individual’s fecal time series and deep sequencing.
Isolating the virulent strains in the laboratory, Xin confirmed that they shared the same metabolic pathways, suggesting these abilities are what made these different strains successful at inducing inflammation. In-lab modeling further confirmed that the most competitive strains could not only induce inflammation, but replicate even better in inflammatory environments. This suggests that even though the flares were the result of E. coli, the strain had figured out how to feed on the resulting alarm response from the body. These results were published in Frontiers in Microbiology.
Clinical insights and a cautionary tale
Last year Xin, and her colleagues, published the final paper in the trilogy in Inflammatory Bowel Disease, winning an editor’s choice award. In this work, Xin compared metagenomics and metabolomics in 129 IBD patients, following 21 over time as they went through ileocolonic resection surgery.
Diversity continued to diminish following surgeries, even 2 years after surgery. Xin says more studies are needed to see how this loss of diversity affects the health of IBD patients who receive surgery, and whether surgical strategies might need to change to achieve a healthier diversity and stability after operations.
After completing her PhD, Xin joined 23andMe as a computational biologist in the therapeutics department, doing human genetics-based drug discovery. “It has definitely been helpful to have worked in a disease setting as a bioinformatician.” Says Xin.
Page updated
Report abuse