Enabling High-Frequency Longitudinal Measurements of Host-Microbe Interactions in Free-Living Humans

Author: Akana, Reid Toshio Keali’i

Year: 2027

Degree: Dissertation (Ph.D.)

Advisor: Ismagilov, Rustem F.

Committee Members: Thomson, Matthew; Mazmanian, Sarkis K.; Leadbetter, Jared R.; Ismagilov, Rustem F.

Option: Bioengineering

Abstract

Many human–microbe interactions that maintain health or contribute to disease change over timescales of hours to days, yet most human studies rely on measurements collected weeks or months apart or only after symptoms develop. Understanding these dynamic interactions requires longitudinal measurements collected at frequencies that match the underlying biology. This thesis develops approaches for obtaining high-frequency measurements from people living in the community and for characterizing microbial abundance, genomic content, and transcriptional activity directly from challenging human samples.

The COVID-19 pandemic created an urgent need to measure SARS-CoV-2 viral dynamics from before infection became detectable through the early stages of disease. I developed the computational and study infrastructure required to prospectively enroll uninfected people at high risk of infection and support and monitor frequent, high-quality at-home sampling at scale, ultimately enabling studies of more than 400 participants and more than 14,600 samples. Daily measurements of saliva, nasal swabs, and throat swabs revealed substantial differences in viral-load dynamics among specimen types, with oral viral loads often rising days before nasal viral loads. These findings showed that measurements from a single specimen type can miss early infection and informed improved diagnostic sampling strategies.

I next developed molecular approaches to enable analogous measurements of human-associated fungi, whose low abundance relative to human nucleic acids makes direct characterization difficult. To measure absolute fungal abundance, I developed Fungal QuantSeq, a targeted sequencing approach that detects and quantifies diverse fungi in human-DNA-rich samples. To characterize fungal genetic potential, I co-developed a fungal Microbial-Enrichment Method (fMEM) that selectively removes human DNA to enable direct fungal metagenomics and recovery of fungal genomes from clinical samples. Finally, to measure fungal gene expression in samples dominated by human RNA, I developed an RNA-enrichment method that removes human RNA while preserving fungal gene-expression profiles, enabling longitudinal, genome-wide measurements of Candida transcription directly from vaginal swabs.

Together, these approaches establish a framework for measuring microbial abundance, genetic potential, and transcriptional activity at frequencies that better match the dynamics of human–microbe interactions. They enable studies of biological transitions that may be missed by sparse or clinic-based sampling and provide tools for investigating how microbes move between colonization and disease in people living normally in the community.