At BEBPA we track implementation of new technologies, analysis approaches and the emergence of new technologies in the potency bioassay world. Over the past years, we have seen a steady increase in the use of frozen-ready-to-use cells, development of reporter gene assays and the increased use of automation. 

What is next? What about AI (Artificial Intelligence)? We hear about Artificial Intelligence (AI) in all aspects of our lives, but what about AI in the lab? Specifically, what does AI look like in the bioassay laboratory? I thought a fun place to find out the uses of AI in the laboratory – would be to look at an AI Overview of the topic! Here is what I found:

AI Overview

Artificial intelligence (AI) is a technology that is being used in laboratories to improve efficiency, accuracy, and diagnostic processes. AI can help with a variety of tasks, including:

• Data analysis: AI can analyze large amounts of data to identify trends and produce insights. 
• Disease detection: AI can help detect diseases by analyzing lab samples and comparing them to a database of other samples. 
• Test interpretation: AI can help interpret laboratory test results. 
• Predictive maintenance: AI can predict when laboratory equipment might need maintenance. 
• Quality control: AI can help with quality control by identifying when results start to drift. 
• Reducing errors: AI can help reduce errors in diagnosis and treatment plans. 
• Automating tasks: AI can automate routine tasks, such as data entry and sample handling. 
• Reducing redundancy: AI can help reduce redundancy by predicting test results from other patient data. 
• Filling staffing shortages: AI could help fill gaps in the laboratory professional shortage by automating tasks that people don’t enjoy doing. 

Digging deeper and specifically thinking about potency bioassays, it is obvious that the tracking and trending of potency bioassays is an often-overlooked area ripe for the use of AI.

Imagine a world, in which all the characteristics of your potency assay, not just selected system suitability is tracked, trended and emailed to you in a weekly report, all without you having to retrieve, manipulate and “massage” the data. This is particularly important for assays which includes any batch of material used as reference material used to release commercial or clinical batches. 

Thinking back over my career as a bioassay consultant, I can think of numerous cases in which we would have more quickly noticed a decreasing potency of clinical reference material if only the data indicating an increasing background, decreasing upper asymptote, or shifting ED₅₀  had been tracked and trended on a routine basis. Often the first indicator of instability was not the assay characteristic we expected and. Therefore, in the days of manual trending data, it was not selected for monitoring.

Although this might seem like pie-in-the-sky thinking, BEBPA is running a session at the upcoming BEBPA US bioassay which starts with case studies about the automation of running assays and ends with two talks about automating the data analysis, trending this data and automating reports. 

The session line-up includes: