Bioinformatic approaches for exploring the genetic and compositional diversities of human-associated microbial worlds

Abstract

Human health is highly affected by the microbiome that we are exposed to, such as indoor microbial communities and our commensals. On the other hand, selected microbes have also participated in human history over thousands of years and have shown great potential in modern bioengineering and pharmaceutical industries. Over the past decades, genomic and metagenomic sequencing has enabled the genetic and compositional diversities of the microbial world to be revealed. However, to promote the industrial values of microbes by utilizing the vast genetic and functional diversity, omic studies with larger population scale and higher resolution are still urgently needed. Similarly, analyses with high resolution on microbial compositions can also help us discover new species or strains with particular industrial or clinical value, and associate health risks with specific organisms for further surveillance and therapy. Thus, during my PhD period, several high-resolution microbial genomic and metagenomic studies were performed, covering the fields of bioengineering, industrial, environmental and clinical microbiology. Four of these projects -two population genomic studies and two metagenomic studies- were presented as separate chapters in this thesis. In Chapter II, I introduce MESSI, a web server developed for the selection of best S. cerevisiae strains and engineering targets for different bioengineering purposes, in regard to baking yeast’s potential as cell factories. As a continuation of this work, in Chapter III, I characterize 36 industrial and natural S. cerevisiae strains under different stress conditions, uncovering the genotypic, metabolomic and phenotypic divergences between them. I suggest platform strains with resistance to multiple or specific stresses and propose potential regulatory modules for different stress conditions. In Chapter IV, I characterize the human skin microbiome after touching metro car handrails in Hong Kong, discovering that the microbiome and resistome were highly influenced by the intraday sampling time. Interestingly, I also captured the potential of cross-border antibiotic resistance gene (ARG) transmission from mainland China. Some methods developed in Chapter III were also introduced for the identification of the strain-level composition and strain-specific pathogenicity. In Chapter V, a metagenomic study identified the gut species associated with insulin sensitivity and ANGPTL4 levels, which prevents fat storage during resistant starch (RS) intervention on overweight/obese mice, serving as potential source for prebiotic approaches to treat obesity. Other related works and on-going works were also summarized in Chapters I and VI. All the aforementioned projects have shown great value of large-scale and high-resolution studies in the exploration and utilization of the diversity of the human-associated microbes and microbiome, or the importance of the strain- and gene-scale resolution for the health risk surveillance.