Single-cell approaches and deep learning to map all stages of fruit fly embryo development

Scientists have constructed the most comprehensive and detailed single-cell map of embryonic development in any animal to date, using the fruit fly as a model organism.

Published in Science, this study, co-led by Eileen Furlong of EMBL and Jay Shendure of the University of Washington, leverages data from over of one million embryonic cells covering all stages of embryonic development and represents a significant advance on several levels. This basic research also helps scientists’ ability to dig deeper into questions such as how mutations lead to different developmental defects. Furthermore, it provides a pathway to understanding the vast non-coding part of our genome that contains most of the disease-associated mutations.

“The basic fact of capturing the entire of embryogenesis – all stages and all cell styles – to get a more complete view of the cellular states and molecular changes that accompany development is a feat in itself,” said Eileen Furlong, genome manager of the EMBL. Biology Unit. “But what really excites me is using deep learning to get a step-by-step view of the molecular changes driving embryonic development, down to the minute.”

Embryo development begins with the fertilization of an egg, followed by a series of cell divisions and decisions that give rise to a highly complex multicellular embryo that can move, eat, smell and interact with its environment. Researchers have been studying this process of embryonic development for more than a hundred years, but it’s only in the past decade that new technologies have allowed scientists to identify the molecular changes that accompany cellular transitions at the human level. single cell.

These single cell studies have generated enormous excitement because they have demonstrated the complexity of cell fates in tissues, even identifying new forms of cells, and revealed their developmental trajectories in addition to the underlying molecular changes. However, attempts to profile the entire development of the embryo at single-cell resolution have been out of reach due to numerous sampling, cost and technological challenges.

In this regard, the fruit fly (Drosophila melanogaster), a preeminent model organism in developmental biology, gene regulation and chromatin biology, has key advantages when it comes to to develop new approaches to address them. Fruit fly embryonic development occurs extremely rapidly in just 20 hours after fertilization all tissues have formed including the brain, intestine and the heart, so the organism can crawl and eat. This, combined with the many discoveries made in fruit flies that have propelled the understanding of how genes work and their products, has encouraged the Furlong laboratory and its collaborators to take up this challenge.

“Our goal was to get a carry on view of all stages of embryogenesis, to capture all the dynamics and changes as an embryo develops, not just at the RNA level but also control elements that regulate this process,” the co-author said. Stefano Secchia, PhD student in the Furlong group.

Preliminary work with ‘enhancers’

In 2018, the Furlong and Shendure groups have shown the feasibility of profiling ‘open’ chromatin at single-cell resolution in embryos and how these regions of DNA often represent active developmental enhancers. “Enhancers” are segments of DNA that act as control switches to turn genes on and off. The data showed which styles of cells in the embryo are using which activators at any given time and how this usage changes over time. Such a map is key to understanding what drives specific areas of embryonic development.

“I was really excited when I saw these results,” Furlong said. “Going beyond RNA to look upstream at these regulatory switches in individual cells was something I didn’t think was possible for a long time.”

Beyond “snapshots”

The 2018 study was cutting edge of technology at the time, profiling approximately 000 000 cells in three different windows of development of the embryo (at the beginning, in the middle and at the end). However, this work has so far only yielded snapshots of cellular diversity and regulation at precise and discrete times. The team therefore explored the potential of using samples from overlapping time windows and, as a proof of principle, applied the idea to a specific lineage – muscle mass.

This then set the stage for dramatic scaling using new technology developed in the Shendure lab. The team’s current work has profiled the open chromatin of nearly a million cells and the RNA of half a million cells at overlapping time points and spanning the entire development of the embryo from the fruit fly.

Using a spell of machine learning, the researchers took advantage of overlapping time points to predict the weather with much higher resolution fine. Co-author Diego Calderon, a postdoctoral researcher at the Shendure lab, trained a neural network to predict the precise developmental time of each cell.

“Even though the collected samples contained embryos of slightly different ages within a 2 or 4 hour time window, this method allows you to zoom in on n’ any part of this embryogenesis timeline on a scale of minutes,” Calderon said. works. We could capture molecular changes that happen very quickly over time, within minutes, which previous researchers had discovered by removing embryos every three minutes.”

To In the future, such an approach would not only save time, but could also serve as a benchmark for usual embryo development to see how things might change in different mutant embryos. This could determine exactly when and in which cell style a mutant phenotype appears, as the researchers have shown in muscle. In other words, this work not only helps understand how development normally occurs, but also opens the door to understanding how different mutations can mess it up.

The new predictive potential that this research portends, based on samples from much larger time windows, could be used as a framework for other model systems. For example, mammalian embryo development, in vitro cell differentiation, or even post-drug treatment in diseased cells, where deviations in sampling times can be engineered to facilitate optimal temporal prediction at resolution of research.

In the future, the team plans to explore the predictive powers of the atlas.

“In Combining all the new tools available to us in single-cell genomics, computer science and genetic engineering, I would like to see if we could predict what happens to individual cell fates in vivo following genetic mutation,” Furlong said. ” … But we are not there yet. However, before this project, I also thought that actual work would not be likely anytime soon.

Related Articles

Back to top button