It’s been a busy few months for the team here at Aigenpulse as we work on releasing some updates to our CytoML Suite. We managed to grab our Chief Product Officer, Satnam Surae, for a quick chat about how automation can help the pharma industry leverage data analytics and advanced machine learning to uncover insights and facilitate decision making.
What are the advantages of using flow cytometry for drug discovery research?
Flow cytometry is one of many analytical techniques used in pharma R&D to develop compounds and it has been used commercially in research and clinical labs for about 40 years now.
In recent years, developments in technology have resulted in the availability of high-throughput flow cytometry approaches and extended its applications in cell-based assays. The high-speed quantitative analysis of cells and particles enables rapid drug molecule screening, and that makes flow cytometry an exciting technology for drug discovery research. Its multiparameter capabilities generate a range of different types of information.
However, this ability to measure so many parameters means that huge quantities of data are being generated all the time and need to be analysed, which often causes a bottleneck in research. With this data having so much potential to help with life-changing therapies, it’s important to be able to extract insights quickly and efficiently. Automated approaches to flow cytometry are the key to capitalising on the technique’s benefits and freeing up researchers’ time so that they can focus on how to act on the insights uncovered.
How can automation help researchers in pharma and biopharma to manage these large quantities of data?
Automation supports the entire drug discovery process, accelerating and cementing the evidence that drug developers must provide, while also increasing the amount of vital new drugs to those in need.
In life sciences, many organisations are only scratching the surface of how to organise, mine and derive value from their data. The core strengths of most researchers lie in their scientific expertise, not in structuring, organising and managing data, so automation offers a range of benefits for organisations and will ultimately help pharmaceutical companies to create effective drugs for patients.
Taking an automated approach to flow cytometry reduces the hands-on processing time for every analysis while retaining control. Automation of data analysis also allows for computational filtering of individual files for low quality data, plus the identification and normalisation of batch effects across larger datasets. It has an extremely high reproducibility rate and consistency is increased during every step, from raw data to results, so valuable insights are easier to extract.
Through using supervised machine-learning algorithms, automated methods can mimic human gating strategies and are more efficient in reducing time-spent. This ultimately results in reduced overall costs.
What tools are available to help?
New dedicated platforms for pharmaceutical R&D data processing are now emerging that tackle researchers’ data challenges. The use of one common platform that aggregates, structures and digitalises workflows across an organisation can unlock the potential of that data, providing easy access and opportunities for greater and more productive cross-department collaborations.
That is what we offer with our CytoML experiment suite: an automated end-to-end process for large numbers of raw files by leveraging machine learning to empower cytometry data processing. Users can integrate population counts identified by manual gating to increase the value of data and allow for cross-project analysis, which provides significant savings in terms of both time and money and tackles the bottlenecks in the research process. CytoML is underpinned by our state of-the-art data intelligence platform, which is designed to expedite the drug discovery and development process.
With the Aigenpulse Platform, we are lowering the entry barrier for customers to digitally transform. The result is a data repository that becomes an organisation’s single source of truth, where all critical information is aggregated, stored and easily accessed.
Could you talk us through some of the updates you and the team have been making to the CytoML Suite in response to this shift to automated workflows?
Flow cytometry data analysis is built upon the principle of gating, a key step necessary to visualise relevant correlations in multiparameter, multi-population data. However, this opens up the risk of individual interpretation in the data, and even automated processes can result in some challenges. Random variation in automated clustering algorithms can lead to inconsistent results meaning that comparing results from automated gating with each other, as well as with traditional manual gating results, can be difficult.
Where researchers need data to support regulatory use cases, guided/semi-automated analysis is key because it is 100% reproducible. However, there is a depth of rich data that underpins the information provided by flow cytometry, which is where unbiased analysis for exploratory use cases can help uncover new insights by finding novel populations or clustering non-intuitive populations together, for instance.
We’ve released CytoML 5.2 to help users perform automated analyses in an unbiased manner for exploratory use cases, including FlowSOM and Phenograph for algorithm-based clustering, and use powerful dimensionality reduction methods such as tSNE and UMAP to visualise connected data. It takes us beyond current capabilities and continues the evolution of this key technique. You can learn more about the updates to CytoML in our recent blog post, here.
If you’d like to free up your scientists’ hands-on time working with data by as much as 50% whilst dramatically improving reproducibility and compliance, then contact us to discuss your data challenges.