An adult human body consists of approximately 37 trillion cells. Not so long ago it was thought that these came in 220 different species. That number, the result of decades of painstakingly looking through microscopes at slides containing chemically stained tissue sections, gave a sense of the distribution of cellular work required to keep a body running.
A feeling, but only a superficial feeling. Tools now exist that are able to look inside cells and crack them open one by one to release their complementary messenger RNA (mRNA), the molecule that carries genetic information from the cell's nucleus to the protein factories. Molecules of mRNA indicate which genes are active, thus revealing the inner nature of a cell. Cells that look similar under a microscope often turn out to be quite diverse. The number of cell types has therefore risen above 5,000.
Leading 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 certain organs or types of tissue. The HCA aims to catalog the whole cabal: identifying and locating all cell types, healthy and diseased, in every human tissue over the course of a lifetime. Its scope even extends to “organoids,” science's clumsy first attempts to grow living simulacra of organs.
According to Sarah Teichmann of the University of Cambridge and Aviv Regev of Genentech, an American biopharmaceutical company that set up the whole thing, they hope to have a first version of the atlas available next year. Their latest progress report has just been published as a series of articles in Nature and a number of sister journals.
As Dr. Teichmann and Dr. Regev note, there are two types of HCA cards. One, conceptually similar to geographers' maps, connects each cell type to a four-dimensional location in the human body (sampling at different stages of life adds the dimension of time to that of space). The other species is less known. These, called manifolds, are normally used by mathematicians to represent multidimensional mathematical hyperspaces. In the case of the HCA, the numerous dimensions in question are not space and time, but rather molecular features, such as mRNA profiles, that characterize different cell types. By plotting different cell types on the same map, manifold graphs increase the understanding of their similarities and differences.
No cell left behind
The geography of the real world also plays a role. From the beginning, Dr. Teichmann and Dr. Regev determined not to oversample parts of the world (Europe, North America and certain 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 demonstrate the scale of the effort. Placentas, the embryonic development of the skeleton, intestinal inflammation and the formation of the thymus (the organ that produces the immune system's T lymphocytes, the cells destroyed by AIDS) are all discussed.
The findings of these studies are groundbreaking. They confirm previous suspicions that some cellular processes involved in cancer tumor formation are involved in the rapid growth of the placenta. They identify genes expressed in developing bone and cartilage cells that can lead to arthritis later in life. By comparing healthy and unhealthy intestines, they show that one source of disease-causing inflammation appears to be intestinal cells accidentally developing into a type normally found in the stomach. And they provide a detailed description of the thymus based on a standardized representation of that organ.
Perhaps the most intriguing paper of all concerns brain-mimicking organoids. Organoids made up of human brain cells, which themselves are derived from laboratory-created stem cells, are something that gives bioethicists the itch. At this point, deprived of the blood supply needed to grow, they are only three or four millimeters in diameter, so they are unlikely to develop any form of consciousness. But some worry that larger versions might.
However, they are useful for research because they allow the study of living human brain tissue without the need to remove it. But they would be even better if the specific types of neurons, in certain versions of them, could be reliably predicted – because neurons make up a large proportion of known cell types, and each has a different job to do.
The HCA will make this easier. A paper coordinated by Barbara Treutlein 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 both identify the neuron types generated in each organoid and determine how closely they resembled their natural equivalents. The results, put together, create a single diagram for such organoids that shows the strengths and weaknesses of the different protocols, and will help plan future research.
In addition to publishing the latest findings from the project members (although the raw data has been online since it was collected), the articles also allow Dr Teichmann and Dr Regev to set out their views on the use of artificial intelligence (AI) to make the atlas a little closer. to a model of how a person works.
Both are computational biologists by training, and it was this background that led them to come up with the HCA in the first place. Without the software underlying the project, which converts data into maps and allows those maps to be interrogated, the project would not exist. But the couple has bigger visions. They were early adopters of basic modeling, a class of AI (like the major language models that have gained popularity in recent years) that feeds on vast amounts of training data to recognize patterns unobservable to humans.
The basic HCA models are not trained on text passages, but on collections of cells. And their goal is not human-like composition, but to create better and more useful maps. Some learn from mRNA data about cell types. Others rely on conventional histology slides and more modern iterations thereof, such as light-sheet imaging, which involves scanning sections through three-dimensional samples. These models are now good enough to be used to annotate the cells in new specimens, to search for similar cells in different specimens, and to discover the gene programs behind certain traits. In the future, they should be able to predict how cell lines will develop and even imagine yet unknown cell varieties. Such models are not only faster than human researchers, but can also perform tasks beyond human capabilities.
The result is a system that can be used (and has been used) not only to improve the atlas, but also to put it into practice. For example, drug manufacturers already use HCA data and models to “virtually” screen potential drugs before experimental testing; to predict side effects by discovering non-target tissues where the gene with which a drug candidate interacts is expressed; and, conversely, to discover opportunities in such non-target tissues to expand the range of therapeutic targets of a drug.
One day, all these efforts could contribute to a human “digital twin,” which would also include basic models of how proteins work (such as AlphaFold, a protein folding model developed by Google DeepMind) and how bodies develop. That day is still far away. But it now seems more likely that it will come to that.
© 2024, The Economist Newspaper Limited. All rights reserved. From The Economist, published under license. Original content can be found at www.economist.com
Catch all Business news, Market news, Breaking news Events and Latest news Updates on DailyExertNews. Download the Mint News App to receive daily market updates.
.Fewer