The network of soil microbial community: its structure and dynamics

概要

博士論文（要約）

The network of soil microbial community: its structure and dynamics
土壌微生物の群集ネットワーク：構造的特徴とダイナミクス

2023 Academic Year

Tohoku University, Graduate school of Life Sciences
Department of Ecological Developmental Adaptability Life Sciences

楊 岱霖（YANG DAILIN）

Soil microbes are incredibly diverse and play important roles in the functioning of the Earth. Through those functions,
microbes bring significant benefits, which are essential for the survival and development of human societies. Better
understanding of the soil microbial community is a key topic of discussion for scientists in an era of constant change in
the earth's environment.
Microbial communities are complex, with an extremely large number of members connected by a variety of
interactions to form the fundamental mode of operation of these important functions. Understanding the member of
microbial communities and the way in which interactions are organized is therefore the most fundamental topic in the
study of soil microbial function. However, in contrast to macroscopic organisms where this information can be obtained
by observation, the extremely small size and large numbers of microbes make it challenging for humans to identify those
interactions.
Methods have been developed to obtain information on interaction by incubating microbial members in pairs.
However, as less than 1% of microbes can be isolated from communities and then incubated individually, it is extremely
challenging to explore microbial communities using this approach. In addition, the interactions identified by incubating
in pairs are not necessarily present in complex natural communities.
The development of high-throughput sequencing technologies has provided humans with a powerful tool to recover
the composition of microbial communities by detecting the genetic material present in the community. However, we still
struggle to explore the interactions between members of these microbes by only use it. Although there are methods
proposed aiming to identify interactions by combining high-throughput sequencing and correlation between members,
such correlation-based interaction studies are less reliable because microbial interactions are essentially causal rather than
simply correlational. Modelling approaches may be useful for identifying causality, but accurate modelling requires a
large amount of basic knowledge about soil microbial communities, which is currently lacking. The first step in soil
microbial ecology research is therefore to integrate existing technologies to reliably identify the interactions of large-scale
microbial communities.
In this study, I combine high-throughput sequencing techniques with the non-linear time series analysis technique
UIC to infer interaction networks in soil microbial communities. Rather than relying on models, the non-linear time series
analysis technique is based on a series of delays in the time series to recover the dynamical system of the whole community,
whereby causal relationships can be effectively identified. I have used this technique for the first time to construct an
interaction network of a large-scale soil microbial community and have built on it to analyze a range of ecological issues.

In Chapter 2, I explore the applicability of the stress gradient hypothesis in soil microbial communities, by applying
a nonlinear time series analysis to the amplicon-based diversity time series data of the soil microbiota cultured under less
stressful (30°C) or more stressful (37°C) temperature conditions, I show the response mechanisms of soil microbial
interaction networks under stressful conditions.
In Chapter 3, I focus on the important members (keystone and indicator species), that is, using the causal inference
theory-based definition of keystone-ness and indicator-ness, I propose a methodology to identify keystone and indicator
species from time-series data and apply it to the experimental microbial communities.
In Chapter 4, I focus on the recovery behavior of microbial communities and identify members that have a significant
influence on this behavior and that may act as controlling factors for community recovery. I found that only a handful of
members were important in recovering community composition, and at different levels of stress, these members have
different identities and different patterns of impact on community recovery.
In summary, my research has deepened human understanding of the existence patterns, response patterns and
stability of soil microbial communities. ...