The developing branching lung in an in vitro culture system
Image Credit: Miram Meziane/Whitehead Institute
Scientific Frontline: Extended "At a Glance" Summary: IRIS Machine Learning Model for Cellular Signaling
The Core Concept: IRIS is a neural network-based machine learning model designed to reconstruct the developmental signaling histories of single cells by analyzing their overall gene activity. It identifies specific chemical communication sequences that guide undifferentiated stem cells to assemble into specialized three-dimensional tissues and organs.
Key Distinction/Mechanism: While researchers previously assumed that the effects of signaling pathways varied too widely across cell types to be mapped universally, this framework demonstrates that each signaling pathway leaves a consistent, unique transcriptional "fingerprint" across diverse cell types. Utilizing transfer learning, IRIS detects these pathway-specific patterns to pinpoint which chemical signals a cell received at different stages of embryonic development, even in previously unencountered cell types or across different species.
Major Frameworks/Components:
- Perturbation Screen Training Atlas: The model was trained on a large experimental dataset measuring how thousands of human embryonic stem cells responded to dozens of combinations of six major signaling pathways across multiple developmental stages.
- Pathway-Specific Transcriptional Fingerprints: Distinctive, conserved patterns of gene expression that reflect the exact multi-step chemical signaling cascades a cell has encountered.
- Cross-Species Transfer Learning: A neural network architecture capable of translating training data from human embryonic stem cells to accurately predict spatiotemporal pathway activations in single cells from mouse embryos during gastrulation—covering heart, gut, muscle, and spinal cord tissue formation.
- Experimental Validation in Lung Morphogenesis: In vivo verification in mouse embryos confirming the model's prediction of specific signaling pathway activations required to drive lung-specific cellular development.
Branch of Science: Developmental Biology, Computational Biology, Machine Learning, Stem Cell Biology, Systems Biology, and Regenerative Medicine.
Future Application: The framework provides a precise roadmap for controlling stem cell fate in vitro, accelerating stem cell engineering and improving the generation of accurate three-dimensional organoids (such as lab-grown lung tissue models). These engineered tissues can be deployed to study complex disease mechanisms—including asthma, lung cancer, and idiopathic pulmonary fibrosis—and to screen targeted pharmacological therapies or design regenerative tissue treatments.
Why It Matters: Traditional methods require researchers to experimentally test every signaling pathway in every individual cell type, a slow and labor-intensive barrier to developmental mapping. By decoding intercellular communication through transfer learning, IRIS allows scientists to comprehensively map the signaling histories of every cell inside a mouse or human embryo at an unprecedented scale and speed.
As an embryo develops from a small cluster of stem cells, those once "blank-slate" cells begin to take on more specialized roles, such as brain, liver, or muscle cells, and organize themselves into three-dimensional structures such as tissues and organs.
The fate of each cell—what type of specialized cell it will become—depends on which genes are turned on or off within the cell. These patterns of gene activity shape the cell's structure and function, enabling it to take on a specific role in the body.
However, this decision is not up to individual cells. They constantly send and receive chemical signals to and from neighboring cells, which helps them determine where they are, what stage of development they are in, and what they should become. These messages spread through multistep sequences called signaling pathways, which translate external signals into specific changes in gene activity within the cell.
For researchers, being able to retrace the sequence of instructions a cell has received would offer a powerful way to understand how tissues develop and how these processes go awry in disease.
Yet this has been difficult to achieve because scientists have long assumed that the effects of signaling pathways vary widely across cell types, meaning they would need to map each pathway separately in each cell type—an arduous and painstaking process.
Now, in a new study led by Massachusetts Institute of Technology (MIT) assistant professor of biology and Whitehead Institute for Biomedical Research member Pulin Li and graduate student Nicholas Hutchins, researchers have discovered that each signaling pathway leaves behind a unique "fingerprint"—a distinctive pattern of gene activity that reflects the particular signals the cell has encountered.
Importantly, these fingerprints are consistent across different cell types for the same signaling pathway, which means that instead of mapping each cell type separately, scientists can reconstruct signaling histories across many cell types using these pathway-specific fingerprints.
This discovery was made possible through a machine learning model called IRIS. This model can detect the fingerprints of different signaling pathways and pinpoint which signals a cell received at different stages of development inside an embryo, even for cell types it has not encountered before.
This artificial intelligence–driven approach marks a major advance over traditional methods, which require researchers to experimentally test every pathway in every possible cell type, and opens up the possibility of comprehensively mapping the signaling histories of every cell inside a mouse or human embryo at an unprecedented scale.
"Think of voice recognition systems like Siri, which are trained mainly in English, but then use that training to help them recognize other languages," says Li. "This is called transfer learning, and this is why IRIS can work across many different cell types."
The researchers' detailed open-access findings, published in the journal Nature Methods on September 8, could accelerate stem cell engineering for regenerative medicine and improve the creation of organoids—miniature, three-dimensional models that mimic real organs—for studying disease mechanisms and testing new drugs. This is because once researchers learn the pattern of signals that drives a stem cell to become a specific cell type, they can recreate those signals to control the fate of stem cells in a laboratory or medical setting.
IRIS is a neural network–based model, which, in essence, is an artificial intelligence–driven system designed to recognize patterns in complex data, similar to how the human brain spots patterns. It examines a cell's overall gene activity and estimates which signaling pathways were likely "on" at specific times during development.
Li and Hutchins trained IRIS on a large experimental dataset that measured how thousands of human embryonic stem cells responded to dozens of combinations of six major signaling pathways at multiple stages of development. This created a comprehensive map, or atlas, of how signaling combinations influence cell behavior.
The team then tested IRIS using single cells from mouse embryos during gastrulation, a stage when cells are rapidly branching into different fates. IRIS could accurately predict when and where specific signaling pathways would activate in cells destined to become part of heart, gut, muscle, and spinal cord tissue. This variation in signaling molecules is what guides cells to form the right structures in the right places.
With this approach, scientists can not only begin to understand the fundamentals of cell-to-cell communication, but also gain a practical roadmap for guiding stem cells into specific, functional cell types in the laboratory.
When the researchers employed IRIS to identify the signals needed to create a cell type critical for lung development, the model predicted that activating a specific signaling pathway would encourage lung-specific development. Experiments in mouse embryos confirmed the model's prediction. By identifying the precise combinations of signals that drive lung cell development, researchers can more reliably generate accurate, laboratory-grown models of lung tissue.
"In these ways, IRIS is helping us decode the language cells use to talk to each other at a much faster rate than we could realistically achieve through experiments," Hutchins says.
These improved models will allow researchers to study diseases such as asthma, lung cancer, and pulmonary fibrosis, in which lung tissue becomes scarred, often without a known cause. They can then use these models to test potential therapies and ultimately design regenerative treatments that can repair the damaged tissue.
Published in journal: Nature Methods
Title: Reconstructing signaling histories of single cells via perturbation screens and transfer learning
Authors: Nicholas T. Hutchins, Miram Meziane, Claire Lu, Maisam Mitalipova, David S. Fischer, and Pulin Li
Source/Credit: Massachusetts Institute of Technology | Shafaq Zia, Whitehead Institute
Edited by: Scientific Frontline
Reference Number: bio101626_01
