Scientists are making a list of every type of cell in our body. mint

An adult human body contains approximately 37 trillion cells. Not long ago it was thought that these came in 220 different types. That number is the result of painstaking decades spent looking through microscopes at slides containing tissue sections stained with chemical stains, realizing the division of cellular labor needed to keep the body running.

A feeling, but only superficial. Tools now exist that are able to look inside cells, breaking them down one by one to release complements of messenger RNA (mRNA), the molecule that carries genetic information from the cell’s nucleus to its protein factories. . mRNA molecules indicate which genes are active, revealing the internal nature of the cell. Cells that look similar under the microscope often turn out to be quite diverse. Thus the number of cell-types has increased above 5,000.

The leader of this histological revolution is the Human Cell Atlas (HCA) Consortium, which was founded in 2016 and currently includes more than 3,600 collaborators in 190 laboratories in 102 countries. Other cell-atlas projects are limited to mapping particular organs or tissue types. HCA aspires to catalog the entire constellation: identifying and locating all cell types, healthy and diseased, in every human tissue during the lifetime. Its scope also extends to “organoids”, science’s first attempt to develop living simulacra of organs.

The first draft of the atlas is expected to be available next year, according to Sarah Teichman of Cambridge University and Aviv Regev of US biopharmaceutical firm Genentech, who created it all. Their latest progress report has just been published as a set of papers in Nature and several of its sister journals.

As Dr. Teachman and Dr. Regev explain, there are two types of HCA maps. One, similar to the concept of geographers’ maps, connects each cell type to a four-dimensional site in the human body (sampling at different stages of life connects the dimension of time to the dimensions of space). Other types are less familiar. These, called manifolds, are commonly used by mathematicians to represent multidimensional mathematical hyperspaces. In the case of HCA, many of the dimensions in question are not space and time, but rather molecular features, such as mRNA profiles, characteristic of different cell types. By plotting different cell types on the same map, charts of manifolds thus enhance the understanding of their similarities and differences.

no cell left behind

Real-world geography also plays a role. From the beginning, Dr. Teichman and Dr. Regev have been determined not to observe the parts of the world (Europe, North America, and parts of Asia) where scientists are concentrated. Instead, they sought participants from all six inhabited continents—a decision that has already been rewarded with insights into the cellular basis of geographic differences in immune responses and breast cancer susceptibility.

The topics of this week’s papers reflect the scope of the effort. Placentas, fetal development of the skeleton, inflammation of the intestine, and formation of the thymus (the organ that produces the immune system’s T-lymphocytes, cells damaged by AIDS) are all discussed.

The findings of these studies break new ground. They confirm earlier suspicions that some of the cellular processes involved in cancerous tumor formation are involved in the rapid growth of the placenta. They identify genes expressed in developing bone and cartilage cells that can cause arthritis in later life. They compare healthy and unhealthy intestines to show that one source of disease-causing inflammation is intestinal cells that mistakenly grow into the type found in the stomach. And they give a detailed description of the thymus based on a standardized representation of that organ.

However, perhaps the most intriguing paper is on brain-mimicking organoids. Organoids made from human brain cells, which are themselves derived from lab-grown stem cells, are something that troubles bioethicists. At the moment, deprived of the blood supply needed to grow, they only reach three or four millimeters long, so are unlikely to develop any form of consciousness. But some people are concerned there could be bigger versions.

However, they are useful for research, as they allow the study of living human brain tissue without the need to remove any. But they would be even better if their particular versions could be reliably predicted for particular types of neurons – because neurons collectively make up the vast majority of known cell types, and each has a different job to do. Is.

HCA will make it easier. A paper coordinated by Barbara Trautlein of the Federal Institute of Technology in Zurich looked at mRNA data from 36 such organoids created using 26 different protocols. The researchers involved were able to identify the neuron types generated in each organoid and determine how closely they resembled their natural counterparts. The results, stitched together, create a single manifold chart for such organoids that shows the strengths and weaknesses of different protocols, and will help plan future research.

In addition to publicizing the project members’ latest findings (although the raw data have been online since they were collected), the papers call on Dr. Teichman and Dr. Regev to use artificial intelligence (AI) to bring the atlas some closer. Also allow you to establish your point of view. A model of how a human being works.

Both are computational biologists by training, and it was this background that led them to first conceive of HCA. Without the software underpinning the project, which transforms data into maps and allows those maps to be interrogated, the project would not exist. But this pair has bigger visions. They were early adopters of foundation models, a class of AI (such as the large language models that have gained prominence in recent years) that feed on large amounts of training data to recognize patterns that are incomprehensible to humans. Does.

The basic model of HCA is trained not on fragments of text, but on collections of cells. And their goal is not to create human-like creations but to create better and more useful maps. Some people learn about cell types from mRNA data. Others rely on traditional histology slides and more modern iterations thereof – such as light-sheet imaging, which scans sections through three-dimensional specimens. These models are now good enough to be used to annotate cells in new samples, discover similar cells in different samples, and discover the gene programs behind particular characteristics. In the future they should be able to predict how cell lineages will evolve and even hypothesize as yet unknown cell varieties. Such models are not only faster than human researchers, but can also perform tasks beyond human capacity.

The result is a system that can (and has) not only been used to enhance Atlas, but can also be harnessed. For example, pharmaceutical companies are already using HCA data and models to “virtually” screen potential drugs before experimental testing; predicting side-effects by exploring non-target tissues where genes that interact with the drug candidate are expressed; and, conversely, to explore opportunities in such non-target tissues to expand the range of the drug’s therapeutic targets.

All this effort may one day contribute to a human “digital twin,” including basic models of how proteins work (like AlphaFold, a protein-folding model developed by Google DeepMind) and how bodies develop. Will be included. That day is still far away. But now the possibility of its arrival seems more likely.

© 2024, The Economist Newspaper Limited. All rights reserved. From The Economist, published under license. Original content can be found at www.economist.com