How AI Could Improve Flow Cytometry Workflows

The use of artificial intelligence (AI) has bloomed in the past few years, becoming a valuable tool in almost every area of life, from content creation and data analysis to answering everyday questions.

Flow cytometry is a versatile and powerful technique that enables the high-throughput and multiparametric evaluation of individual cells in a sample. As a result, flow cytometry is extensively utilized across various scientific fields, such as immunology, oncology, and veterinary research.

Over time, flow cytometry has advanced significantly to allow the use of bigger and more intricate panels. This is advantageous in that more information can now be obtained from a single sample; however, it has also added an extra level of complexity to the planning, execution, and analysis of the experiment.

In this blog, we discuss how AI could be used to simplify and streamline the execution of flow cytometry experiments.

Using AI to Simplify Flow Cytometry Panel Design

Whether you are a beginner or an expert, multicolor panel design can be a time-consuming and convoluted part of the flow cytometry workflow, and bigger panels come with more aspects to consider.

The first step in panel design is selecting the appropriate markers to identify and characterize the cell populations you are interested in, while weeding out those you are not. This often involves an in-depth literature search of previously published panels, which can be a lengthy process.

Here, AI could be a useful tool to sift through the abundance of available information and pull out what is valuable for you. This, of course, would likely still require a sense check to ensure that the presented information is accurate and relevant to your experimental conditions, but could save you time in the long run.

Effective panel design relies on adherence to certain best practice rules. For example, overlapping spectra between fluorophores should be avoided as much as possible, and dim fluorophores should be paired with highly expressed targets while bright fluorophores should be paired with targets with low expression.

Manually building a panel while keeping multiple rules in mind can be overwhelming. To help manage this, many panel builder tools exist with the additional incorporation of AI algorithms that can detect suitable fluorophores based on reduced spillover and spread, antigen density, and compatibility with your instrument.

Bio-Rad’s Panel Builder Tool simplifies the complex panel design process by incorporating an AI-powered platform that allows you to apply your settings and preferences and generate an optimized panel with ease.

Using AI for Flow Cytometry Data Analysis

As flow cytometry technology has progressed, the amount of data that can be collected from a single experiment has increased dramatically. For the researcher, this can mean trawling through mountains of data to find significant changes, a time-consuming and often tedious process.  

It should come as no surprise, therefore, that currently one of the most extensive uses of AI in flow cytometry is in data analysis.

One of the core principles in flow cytometry analysis is gating. Cell populations and subsets are identified by common characteristics, such as forward scatter (FSC), side scatter (SSC), or the expression of certain markers, and distinguished by drawing boundaries (known as a gate) around the cells on the plot.

Traditionally, this is done manually, and the specific placement of gates can be subjective to the user.

Now, many flow cytometry analysis tools integrate AI to achieve automatic gating, saving the user time and enhancing the reliability and reproducibility of results (Yue et al. 2025).

 AI-Assisted Flow Cytometry in Clinical Diagnostics

Reducing user-based differences in analysis is particularly important in clinics when flow cytometry is used as a diagnostic technique, as any variation could impact accurate diagnosis.  

With this in mind, Lu et al. (2024) developed and validated a flow cytometry workflow that used AI to identify immunological disorders from patient blood samples. The AI model they developed was able to accurately distinguish T cells, B cells, and natural killer (NK) cells, and could even identify their subsets, including CD4+ T helper cells, CD8+ cytotoxic T cells, double-negative T cells, and class-switched B cells.

Not only was the AI analysis accurate, but it also saved time, taking less than 5 minutes per case compared to the conventional analysis workflow, which takes 10–20 minutes.

This is not the only example of AI being used to aid in the diagnosis of patient samples by flow cytometry. Various studies have examined its efficacy in other diseases, such as acute leukemia, Hodgkin lymphoma, and B-cell neoplasms, and have similarly noted accuracy and increased time savings (Simonson et al. 2021, Zhao et al. 2020, Zhong et al. 2022).

The Future of AI in Flow Cytometry

With the ever-expanding panel size capacity of flow cytometry experiments, the integration of AI into existing workflows is becoming more and more valuable. Streamlining the workflow, from planning to analysis, offers users precious time to focus on other areas of their work.

While the potential benefits of using AI in flow cytometry for both research and diagnostic purposes are evident, several challenges also exist that will need consideration. For example, using AI in diagnostics will require training AI tools with the right data from patients of all different demographics to prevent bias.

Ultimately, AI is unlikely to replace the expertise of flow cytometry users, but it can help improve consistency and free up time for scientific interpretation. As these tools continue to develop, they may become an increasingly valuable support for researchers and clinicians looking to generate reliable data from complex experiments.