Advancing understanding of the evolution of key bacterial and fungal genes in microbial communities through metagenomic assembly optimisation and context-aware graph algorithms

Microbes underpin the functioning of every ecosystem on the planet. Metagenomic sequencing, in which we determine the DNA of all the microbes in a sample at once, has been a vital tool in understanding what microbes do, how they interact with one another, and how this impacts us. By looking at the interacting ecosystem of microbes in environments such as the oceans, farms, and hospitals we can try to answer questions relevant to humanity: How do microbes become resistant to antibiotics? How do microbes respond to climate change? What leads to the emergence of new infectious diseases? Comparing DNA from different microbes has taught us that DNA can sometimes transfer between unrelated microbes in a process known as lateral gene transfer. We've also learnt that the DNA that surrounds a particular gene plays an important role in what the gene does, how it evolves, and how likely it is to be transferred in this way. Therefore, to understand how functions like antibiotic resistance evolve and spread in microbial communities it is important to be able to identify which microbe a piece of DNA came from, which gene in that DNA leads to that function, and what DNA is next to that gene. Recently, two advances in metagenomic analysis may help solve these problems: methods to group DNA sequences from the same microbe together (metagenome-assembled genomes) and methods looking directly at the network formed when we assemble DNA fragments into longer sequences (sequence graphs). Unfortunately, most tools to do these things have significant shortcomings. Grouping methods tend to focus only on bacteria despite other microbes such as fungi and amoeba playing important roles in the function and evolution of microbial communities. They also perform poorly for some of the most important types of DNA if we want to study lateral gene transfer: mobile genetic elements. On the other hand, the methods for analysing sequence graphs are very computationally demanding to run and prone to giving out incorrect results. Therefore, this proposal encompasses two complementary projects seeking to address these shortcomings. Firstly, we will develop ways to automatically identify the best possible combination of tools and settings to correctly group DNA from mobile genetic elements and microbes like fungi and amoeba. We will then use this to try to better understand which bits of DNA control things a microbial community can do e.g., resist antibiotics or digest plastics. Secondly, we will develop ways of identifying cases of lateral gene transfer using sequence graphs. By mapping these patterns we may be able to predict when and why a gene is transferred between microbes. Together, these results will teach us how to optimise our current tools and how to learn more from the data we've already collected. This will enable us to better understand how microbes interact and how microbial functions evolve and spread with implications for medicine, agriculture, engineering, and ecology.