in

Artificial intelligence to figure out how bacteria infect

Many disease-causing bacteria use a molecular ‘syringe’ to inject a multitude of their proteins, called effectors, in intestinal cells, thereby blocking key immune responses.

Now, an international team of scientists from the United Kingdom, Israel and Spain, from where the Polytechnic University of Madrid (UPM) participates, have joined forces to analyze all these protein molecules together, combining laboratory experiments and research tools. artificial intelligence (AI).

With laboratory experiments with mice and artificial intelligence algorithms, it has been discovered that the protein effectors of bacteria work as a network, which allows them great flexibility to maintain their pathogenicity.

The authors, who publish their study in the journal Science, have used 100 variants of the bacterium Citrobacter rodentium from mouse to model the function of all effectors in vivo. In this way they have discovered that they work together as a network, which allows the microbe great flexibility to bypass the immune system and maintain its pathogenicity.

The artificial intelligence platform correctly predicted alternative network colonization results from in vivo data. The UPM researchers, the AI ​​professor Alfonso Rodríguez-Patón and the doctoral student Elena Nunez Berrueco used the data collected in the laboratory to construct the machine learning model.

Novel AI techniques

The number of possible combinations of effectors exceeds one billion, so studying all the variants would take more than a thousand years of experimental research. That’s where the AI ​​comes into play to change the rules and allow this complex mechanism to be deciphered. The algorithm developed at the UPM is capable of predicting the infective capacity of any variant after learning the patterns of the 100 laboratory experiments.

The algorithm developed at the UPM is capable of predicting the infective capacity of any variant of a bacterium that infects mice after learning the patterns of 100 laboratory experiments

“By studying such a complex biological system, AI is able to see what is not evident before our eyes,” explains Núñez. Predictions help us identify the most relevant effector combinations and thus save time and resources. We can use this model to predict whether a new strain, with a combination of effectors different from those studied, can manipulate our cells and the way he does it. “

The algorithm It is inspired by the networks of artificial neurons, but incorporates knowledge about the target targets of the effectors. The architecture of this network has a particularity: instead of being generic, it has the same shape as the network of biological interactions of the effectors with the components of our cells. This has made it possible to train the network with a very small number of cases, also giving rise to a model with interpretable results (the so-called explainable AI).

With the help of the model, scientists have been able to direct the following experiments towards the most interesting variants. Thus, they have been able to discover small groups of these molecules that are essential. This means that when they are removed or blocked, the bacteria do not infect, posing a promising target for future treatments to help defeat these clever invaders.

When some molecules are eliminated or blocked, the bacteria do not infect, which is a promising therapeutic target

New therapies

In fact, the authors also observed that the host mouse is capable of adaptation, being able to circumvent the obstacles erected by the different effector networks and activate complementary immune responses that eliminate the pathogen and induce protective immunity.

Rodríguez-Patón concludes: “Artificial intelligence appears once again as a disruptive technology, in this case in the field of microbiology. This interdisciplinary research has required us to develop novel AI techniques to unravel the complex network of molecular signals that bacteria use to infect us. The results obtained are very satisfactory, so we will continue to collaborate with the group of Gad Frankel –One of the main authors– at Imperial College London in future research “.

Reference:

David Ruano-Gallego, Gad Frankel et al. “Deconstructing a type III secretion system effector network unravels the inherent robustness and plasticity in pathogenesis and immunity”. Science, 2021.

Rights: Creative Commons.

‘Hotel Transylvania: Transformation’ launches trailer expanding Dracula’s family with a “new monster”

How a healthy diet is key to building immunity against Covid