Back

Macroscopic laws for a microscopic world

New research gives unexpected insights on human microbiome
Macroscopic laws for a microscopic world

Inside the human body lives an ecosystem of organisms, perfectly coexisting in a delicate equilibrium of mutual dependence and each contributing to the complex mechanism of life. Scientists have long been studying microbial communities from all different point of views, from medicine to biology, in search of a complete description of these complex systems.

A new open-access paper published in Science Advances by Silvia Zaoli, a postdoctoral fellow in ICTP's Quantitative Life Sciences section and her colleague, researcher Jacopo Grilli, approaches the study of a particular microbe community, that of the human gut microbiome, through the lens of statistical physics.

The paper investigates how the microbial communities that live in humans' guts change in time, and how they differ between two different individuals. "There were two main questions that we wanted to answer," says Zaoli. "The first was, if I look at a person's gut microbiome, how does it change in time? And secondly, how is this community different from the community that lives in another person's gut?"

To answer these questions, Zaoli and Grilli used data sets already available in literature, describing time series of gut microbial communities from 14 human hosts, sampled daily for a few months, up to over a year. This data represents the species of microbes in every host and their concentrations, and are collected through DNA sequencing of each sample.

"This allowed us to see which microbial species were present in these communities every day and then follow them in time and across individuals," says Zaoli. "From this analysis we highlighted some trends, some patterns that emerged from the data, and from the model that reproduces and describes them."

The two ICTP researchers worked with a massive amount of data, given that an average human being hosts more than 40 trillion, or 1.5 kilos, of bacteria. "If we look at the composition of the bacteria communities inside the body, we can imagine that inside our gut there is a tropical forest, with many different species, a great diversity of shapes and functions," says Grilli. "This diversity changes over time: but while a tropical forest changes on the scale of decades, the ecosystem in our gut changes on the scale of just a few days."

To navigate this complexity, the authors applied the methods of statistical mechanics, describing the system with macroscopic variables, that is, considering statistical properties rather than particular ones. This means that, instead of looking at single entities in the community and studying every interaction between each of them, they looked at the distribution of abundances of species, and probabilistically estimated the way they change in time. The researchers were surprised to discover that two individuals' gut microbiomes are much more similar than expected.

"Traditionally, it was thought that each person had a characteristic microbiome composition, probably dependent on genetic factors or their diet," says Zaoli. "We found that this is not really the case, but rather it looks like the composition of the community inside a person's gut changes in time exploring all possible states, or at least many of them. So, this community is less specific of the individual than what was believed before."

Another way of seeing this is that if the model is left to evolve long enough, the difference between two different individual's microbiomes would be statistically similar to the difference in a person's own community in two different moments in time.

"In our model, changes happen on two time scales," says Grilli. "One is short-term, with relatively small changes that fluctuate around the average composition that we can observe from the data. Then, there is a long-term time scale, of about six to 18 months. On this longer time scale, we observed that every now and then the microbic community "jumps" suddenly, makes a sudden variation around the average of its composition."

The main new element of this research is the fact that the model makes it possible to extend the time range of the system analysis. Previous studies were based on data sets spanning only up to a few months of evolution. With this new tool the researchers could "stretch" the time of observation - and prediction - to over a year, thus making it possible to observe these abrupt changes. "Even though we don't know what exactly causes these jumps," says Zaoli, "we could conclude that after a long-enough time a person's microbiome would actually change much more than we expected, and this change is comparable to the difference between the microbial community from two different people."

The advantage of using methods from statistical physics to describe complex systems such as the human microbiome is in the fact that just a few parameters in the model summarize all the different mechanisms, both internal and external, that together cause an overall observable change. "And from the - few - data that are available until now," comments Grilli, "our model seems to be working well."

Understanding the dynamics of these communities over time, and how their composition differ in different people, could for example be helpful in medicine to design new drugs and probiotics. "Our work doesn't have an immediate medical application, but advances the knowledge on these communities' dynamics," says Zaoli. "The more you can understand how they change in time, how they differ between individuals, how they respond to external stimuli, the easier it is to diagnose any unhealthy condition, since many studies seem to prove that our gut microbiome is linked to our health in many aspects."

On the other hand, a deeper understanding of a system with a high complexity of species such as the gut microbiome can help to find new ways to select few, important properties and parametrize the relevant features. "Developing tools that allow the understanding of the dynamics of such complex systems could be useful also in other contexts," says Grilli. "For example, designing better prediction metrics for how highly diverse communities change over time could be of huge importance during this time of biodiversity loss and climate change in which we currently live."

 

 

--- Marina Menga

Publishing Date