Digitally improving FACS workflows the ‘Aigenpulse Way’

Above, Carlos del Ojo Elias – Senior Bioinformatician of the Software Engineering A-Team at Aigenpulse!

In this post,we’re going to talk about how to address the challenges of automating an entire FACS workflow, while maintaining the quality and tractability of the process. This post is for you if you create, manage or interpret FACS data. 

The Complexity of FACS Data 

Forward-scatter, side-scatter, and fluorescent data from Fluorescence-activated cell sorting (FACS) creates large amounts of data for analysis.  However, currently much of this analysis is done manually using flat files, Excel spreadsheets and complex/uncontrolled R or python scripts, which leads to inefficiencies and inaccuracies and is subject to human error.  Our goal is to make this process more efficient and to reduce the time and effort on the scientists’ side by automating it. 

In this article, we’ll describe the steps that our Senior Bioinformatician, Carlos del Ojo Elias, had to take to digitally improve the FACS data analysis workflow for our clients.  Carlos is working on both the front end and back end of the system in order to ensure a seamless user experience. 

Creating the Solution 

 In order to reduce manual handling and, therefore, inefficiencies of FACS data, Carlos needed to understand how scientists work with the data, what conclusions they extract and how he could improve the workflow. 

The main challenge was to automate the entire workflow while maintaining its quality, but also preserving the traceability of the process. This was compounded by the fact that there are no standard ways of handling and setting up samples and parameters. 

The first step when running a FACS experiment is to specify the parameters for gating, which is a notoriously subjective process, and some would say a black art! The main challenges with manual gating is reproducibility, as well as the fact that it is labour-intensive, and to address these issues, a number of automated gating algorithms (that make use of clustering) have been developed in recent years. Carlos’ task was to implement several clustering algorithms so that gating could be automated and those parameters used in the gating process stored and re-used in the future, thus improving reproducibility and comparability between datasets. By executing a range of clustering algorithms, and viewing the results as they are generated, the user can better select a satisfactory outcome for their FACS data processing.

Who is Carlos and why was he part of the team for this important development?

Carlos has a background in computer science with a Masters in Bioinformatics. His knowledge of biology has given him the ability to understand biological problems with greater detail than computer scientists without a biological background. This enables him to anticipate what scientists should expect from the solution. Carlos’ interest in biology and in building solutions for scientists is driven by his experience in bridging the gap between data science, software engineering and research. He can ask the pertinent questions which get quickly to the heart of the matter and so positively inform the development process, enabling him to avoid potential pitfalls, side effects or compounded problems further down the line.

The End Result 

Digitally improving the FACS workflow can significantly reduce inefficiencies, inaccuracies and human error in generating insights and drawing conclusions from the data. This was achieved by consulting FACS experts who truly understand the scientific challenges to ensure the final product meets and exceeds their needs. As such, we’re excited to be releasing this new module to the market in the near future and making scientists’ lives easier! 

Aigenpulse Automated FACS Example Results

Bit About Us 

 Aigenpulse is a state-of-the-art software platform focused on accelerating drug discovery and development by using advanced data technologies and machine learning to address three key data problem areas – re-use, efficiency and quality.  Researchers use the platform to structure, share and analyse their research data and to scale repetitive tasks, eliminating the frustrations of dealing with Big Data & expediting discovery and development of better targets and candidates.  Our mission is to combine machine learning and human expertise to facilitate the efficient creation of better drugs and diagnostics, making them available to more people.