Cell Culture: Implementing robotics and artificial intelligence

An automated platform for cell culture combines robotics and artificial intelligence to optimize cell culture protocols and reliably produce specific cell types that could be used for regenerative medicine treatments.
  1. Sujith Sebastian  Is a corresponding author
  1. Clinical Biotechnology Centre, Cellular and Molecular Therapies, NHS Blood and Transplant, United Kingdom

Most animals, including humans, are made up of different organs and tissues formed by specialized cells that perform specific roles. When tissues degenerate or become damaged, the affected cells have to be replaced so the tissues can keep performing their roles. This regenerative potential exists thanks to populations of stem cells in each tissue, which can divide to produce more stem cells – maintaining a constant pool of stem cells for repair – or differentiate into specialized cells to replace damaged cells.

The division and differentiation of stem cells needs to be in balance: if too many stem cells differentiate, the pool of stem cells could become exhausted, but if the stem cells divide uncontrollably this could lead to cancer. However, this balance often fails with age, or due to environmental or genetic reasons. One of the goals of regenerative medicine is to produce differentiated cells in the laboratory that can be used to heal tissues when they have lost the ability to regenerate for themselves (Pizevska et al., 2022). These treatments could have many potential long-term health benefits, including extending life expectancy, and research into these treatments is increasing both in the laboratory and in the clinic.

In order to use cells differentiated in the laboratory in a clinical setting, it is essential that the protocols used to produce the desired cell types reproducibly, and in high enough numbers. This is difficult to do because differentiating cells are highly sensitive to stimuli in their environments, meaning that their culture conditions have to be carefully controlled (Figure 1). Human operators introduce an added layer of variability that is difficult to control for, since each person does cell culture a little differently, and the protocols used to manufacture regenerative medicine products are often complex (Sebastian et al., 2019). Additionally, the requirements that cells and cell-derived products need to meet to be used in the clinic are changing rapidly, and the tests used to assess these requirements can also propagate variability (Simon et al., 2016). As a result of these issues, the quality of the cells produced for regenerative medicine treatments is often unreliable.

Comparison of manual and automated manufacturing of cell therapy products.

(A) Schematic of the development lifecycle of cell therapy products. Cells (left) undergo processing in the laboratory (center), where they are cultured under different conditions, often requiring several complex cell culture steps, to develop safe and efficacious therapy product (right). (B) The processing step of therapy product development can be done manually (left) or automatically (right). Manual processing has several factors (center left) that increase variation in the culture conditions. Causes of variation can be environmental factors, human factors or equipment and technology factors (center), and they lead to manufacturing being less consistent. Automated processing can correct the issues that manual processing introduces (center right).

In recent years, advanced imaging techniques and analysis have improved the reproducibility of the cell culture protocols used for the manufacture of cells in regenerative medicine (Imai et al., 2018). This suggests that further minimizing human error and variability could also greatly improve the manufacturing process. One way to do this would be to mechanize and automate the cell culture procedures. Now, in eLife, Yosuke Ozawa, Koichi Takahashi, Tohru Natsume, and colleagues (who are based at various institutions in Japan) – with Genki Kanda and Taku Tsuzuki as first authors – report on LabDroid Maholo, a robotics and artificial intelligence-based cell processing station that can reliably carry out complex manual cell culture procedures typically performed by trained scientists (Kanda et al., 2022).

Replacing a difficult and repetitive manual process with a machine is a critical step in reducing the variability that arises in cell culture protocols. Indeed, it has previously been shown that automating methods using robotics is an efficient way to reduce human error and operator-dependent variability in the laboratory (Archibald et al., 2016). Kanda et al. now demonstrate this principle in the context of cell culture by combining robotics and artificial intelligence to automate complex cell culture protocols.

The LabDroid Maholo System contains highly accurate and programmable robotic arms to manipulate the cell culture conditions, which allowed the robot to culture stem cells under different conditions. In this case, the LabDroid Maholo System was programmed to differentiate stem cells into retinal pigment epithelial cells. The resulting cells were imaged using the LabDroid Maholo’s integrated imaging system, and computer algorithms and statistical analyses were applied to analyze and quantify the level of cell differentiation against the given conditions using the captured images.

The data obtained from these images was then fed into a mathematical experimental design algorithm that uses a batch Bayesian optimization. This allowed the LabDroid Maholo System to determine which factors in the cell culture procedure needed to be optimized to produce more retinal pigment epithelial cells. Using this method allowed Kanda et al. to reduce millions of potential combinations of cell culture parameters to a more manageable number. Briefly, the batch Bayesian optimization used by Kanda et al. first compares the cell culture conditions against the obtained rate of differentiation, and then decides, with the help of artificial intelligence, whether to include further optimization steps. In this case, three successive optimization steps successfully identified the best cell culture conditions to repeatably differentiate stem cells into retinal pigment cells of adequate quality to be used in cell therapy research.

Kanda et al. provide one of the best examples of how to achieve process control and advanced optimization using a combined robotics and artificial intelligence system. The ability to characterize the cultured cells and establish an optimal cell culture protocol based on the data removes human intervention for these steps, reducing variability. Although the platform will need further improvements to provide end-to-end functionality, it is a step forward towards the reproducible manufacturing of cells and cell-derived products for regenerative treatments.


Article and author information

Author details

  1. Sujith Sebastian

    Sujith Sebastian is in the Clinical Biotechnology Centre, Cellular and Molecular Therapies, NHS Blood and Transplant, Bristol, United Kingdom

    For correspondence
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8952-6003

Publication history

  1. Version of Record published: July 20, 2022 (version 1)


© 2022, Sebastian

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.


  • 2,614
    Page views
  • 264
  • 0

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Sujith Sebastian
Cell Culture: Implementing robotics and artificial intelligence
eLife 11:e80609.
  1. Further reading

Further reading

    1. Computational and Systems Biology
    Swann Floc'hlay, Ramya Balaji ... Stein Aerts
    Research Article Updated

    Wound response programs are often activated during neoplastic growth in tumors. In both wound repair and tumor growth, cells respond to acute stress and balance the activation of multiple programs, including apoptosis, proliferation, and cell migration. Central to those responses are the activation of the JNK/MAPK and JAK/STAT signaling pathways. Yet, to what extent these signaling cascades interact at the cis-regulatory level and how they orchestrate different regulatory and phenotypic responses is still unclear. Here, we aim to characterize the regulatory states that emerge and cooperate in the wound response, using the Drosophila melanogaster wing disc as a model system, and compare these with cancer cell states induced by rasV12scrib-/- in the eye disc. We used single-cell multiome profiling to derive enhancer gene regulatory networks (eGRNs) by integrating chromatin accessibility and gene expression signals. We identify a ‘proliferative’ eGRN, active in the majority of wounded cells and controlled by AP-1 and STAT. In a smaller, but distinct population of wound cells, a ‘senescent’ eGRN is activated and driven by C/EBP-like transcription factors (Irbp18, Xrp1, Slow border, and Vrille) and Scalloped. These two eGRN signatures are found to be active in tumor cells at both gene expression and chromatin accessibility levels. Our single-cell multiome and eGRNs resource offers an in-depth characterization of the senescence markers, together with a new perspective on the shared gene regulatory programs acting during wound response and oncogenesis.

    1. Cancer Biology
    2. Computational and Systems Biology
    Xiangkun Wu, Hong Yan ... Li Liang
    Research Article

    Colorectal cancer (CRC) remains a challenging and deadly disease with high tumor microenvironment (TME) heterogeneity. Using an integrative multi-omics analysis and artificial intelligence-enabled spatial analysis of whole-slide images, we performed a comprehensive characterization of TME in colorectal cancer (CCCRC). CRC samples were classified into four CCCRC subtypes with distinct TME features, namely, C1 as the proliferative subtype with low immunogenicity; C2 as the immunosuppressed subtype with the terminally exhausted immune characteristics; C3 as the immune-excluded subtype with the distinct upregulation of stromal components and a lack of T cell infiltration in the tumor core; and C4 as the immunomodulatory subtype with the remarkable upregulation of anti-tumor immune components. The four CCCRC subtypes had distinct histopathologic and molecular characteristics, therapeutic efficacy, and prognosis. We found that the C1 subtype may be suitable for chemotherapy and cetuximab, the C2 subtype may benefit from a combination of chemotherapy and bevacizumab, the C3 subtype has increased sensitivity to the WNT pathway inhibitor WIKI4, and the C4 subtype is a potential candidate for immune checkpoint blockade treatment. Importantly, we established a simple gene classifier for accurate identification of each CCCRC subtype. Collectively our integrative analysis ultimately established a holistic framework to thoroughly dissect the TME of CRC, and the CCCRC classification system with high biological interpretability may contribute to biomarker discovery and future clinical trial design.