Machine learning successfully replicates cell architecture

A new study published in the journal Cell Systems on November 20, 2019, reports the use of machine learning to help form complex cell architectures from pluripotent stem cells, a sophisticated technology that could solve multiple issues that currently hampers the production of artificial tissues and organs.

stem cellsImage Credit: nobeastsofierce /

Medical scientists faced with irreparably damaged organs have long wanted to know how to stimulate their regeneration or to replace them with new ones, to prolong survival and to provide improved quality of life.

Another equally important area of research involves creating artificial tissues which are identical to those in the body, in order to help understand how disease processes evolve and which drugs can be used to treat such disorders. This means that scientists must know how to direct the development of stem cells in the desired pattern to form multiple tissues in the right way.

Stem cells

Pluripotent ('capable of multiple tasks') stem cells are cells that can divide indefinitely or can develop into any of the three germ layers found in the early embryo. Germ layers are the fundamental tissues that give rise to all the different types of cells and tissues found in the adult organism. Given this property, stem cells are potentially able to recreate any tissue or organ found in the body. However, they need to be directed to form the right patterns as they mature. Such signals and organizers are found at the right times in the natural environment, allowing the development of a mature organism.

However, replicating this spatial organization of natural tissues in the laboratory environment is a challenging task, and among the biggest obstacles to crossing the bridge between stem cells and mature functional organs. One current approach is 3D printing, which creates a biocompatible matrix over which stem cells can divide to create the desired shape and complexity. However, recently, it has been found that stem cells can migrate from the spot where they were laid down, leading to tissue aberrations and loss of function.

The new model

Gladstone Institutes researchers have made use of a computational approach to help them in this all-important task of directing the development of stem cells into the right spatial arrangement.

We've shown how we can leverage the intrinsic ability of stem cells to organize. This gives us a new way of engineering tissues, rather than a printing approach where you try to physically force cells into a specific configuration."

Todd McDevitt, corresponding author

The stem cells used in this study are pluripotent stem cells (iPSCs), which are derived from mature adult somatic cells engineered to regain stem cell characteristics such as self-renewal (the ability to keep dividing into its clones indefinitely), and pluripotency. These have already been used to recreate models of almost all kinds of cell types, such as brain cells, heart cells, and kidney cells. Such cells are now being used in cultures in order to find out how diseases occur and to use as transplants in patients. But the fact remains that both structurally, functionally and biologically, these loose disordered cell clusters that comprise a mix of several types of cells at most, aren't the same as a fully developed, complex, properly arranged organ.

Using gene editing to arrange stem cells

Earlier research showed that knocking out (inhibiting) two genes called ROCK1 and CDH1 caused stem cells in a culture medium to grow in a different arrangement. This gave rise to the idea that they could find out how exactly different cell arrangements could be brought about by allowing each of these genes to be expressed at different levels at different time points. The problem is the huge complexity involved in creating such a model. The scientists found they would have to take too many factors into account – the time of inhibition of each gene, the degree of knockdown, the length of the cell growth time, the number of knockdown cells in proportion to the whole culture – to be cost-effective.

At this point, they turned to the computational field for a better, more accurate, and faster way of predicting possibilities with each knockdown level.

The researchers made use of the versatile and powerful CRISPR-Cas9 to edit these genes at any time, in response to the addition of a drug to the iPSC culture. They also added fluorescent markers that would light up cells in different hues when they stopped expressing either of these genes. All this made it easier to find out how cell arrangements could change with genetic modifications.

Enter computational biology

Initially, the biologists performed quite a few experiments using their laboratory platform, with different doses of the block-inducing drug and using CRISPR-Cas9 at different time points of the process. This was to provide the fundamental data for the machine to begin with. Once these results were fed into the computer's machine-learning program, it started to look for patterns within the provided set of data.

This program makes use of available patterns to simulate or replicate them. In this case, it replicated the patterns it saw in the biological data, and thus predicted how the stem cell arrangement would change when gene inhibition and timing alterations were introduced. It also told the scientists the parameters to recreate in order to bring about the same results.

Using this model, the scientists started to explore the program's ability to generate completely novel patterns, for example, a bull' s-eye target.

The power of this model is that it can generate thousands of data points simulating things that it could take months for me to do in a lab."

Ashley Libby, study author

Being able to simulate thousands of stem cell patterns based on different starting conditions, the researchers were now on the road to defining exactly what they needed to do in order to arrange cells the way they wanted them – such as how much of which drug to add where, at what point, and in which manner. Rather like a recipe to modulate iPSC division and maturation, the researchers learned how and when to knock down CDH1 and ROCK1.

The proof

Now it was time for the final test. The researchers put the suggested protocol for a bull's eye arrangement into practice, and it worked. They saw the stem cells dividing and arranging themselves into concentric circles, just as they had hoped. For the first time, scientists were able to make stem cells do what they wanted them to do without using physical force.

Another researcher, Bruce Conklin, said, "I was just blown away when I first saw the results." He adds that it is every biologist's dream to be able to model cell behavior, and the current work advances our ability towards that goal very significantly.

The future

The research team plans to extend the computational model to include other genes that affect cell arrangement, thus providing for even more configuration possibilities. And of course, they want to start designing a 3D cell system architecture, beyond the 2D arrangements they are now able to create.

McDevitt sums up his team's achievement: "We're now on the path to truly engineering multicellular organization, which is the precursor to engineering organs. When we can create human organs in the lab, we can use them to study aspects of biology and disease that we wouldn't otherwise be able to."

Journal reference:

Libby, A.R.G. et al. (2019). Automated Design of Pluripotent Stem Cell Self-Organization. Cell Systems. DOI:

Dr. Liji Thomas

Written by

Dr. Liji Thomas

Dr. Liji Thomas is an OB-GYN, who graduated from the Government Medical College, University of Calicut, Kerala, in 2001. Liji practiced as a full-time consultant in obstetrics/gynecology in a private hospital for a few years following her graduation. She has counseled hundreds of patients facing issues from pregnancy-related problems and infertility, and has been in charge of over 2,000 deliveries, striving always to achieve a normal delivery rather than operative.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Thomas, Liji. (2019, November 21). Machine learning successfully replicates cell architecture. News-Medical. Retrieved on September 28, 2023 from

  • MLA

    Thomas, Liji. "Machine learning successfully replicates cell architecture". News-Medical. 28 September 2023. <>.

  • Chicago

    Thomas, Liji. "Machine learning successfully replicates cell architecture". News-Medical. (accessed September 28, 2023).

  • Harvard

    Thomas, Liji. 2019. Machine learning successfully replicates cell architecture. News-Medical, viewed 28 September 2023,


  1. Atakan Cetinsoy Atakan Cetinsoy United States says:

    Given the level of complexity of biological systems, it's not surprising that only Machine Learning models can make sense of the underlying processes.
    Atakan Cetinsoy
    BigML - Machine Learning made easy and beautiful for everyone

The opinions expressed here are the views of the writer and do not necessarily reflect the views and opinions of News Medical.
Post a new comment
You might also like...
Mapping the human body one cell at a time: New study reveals the intricate relationship between cell size and count