client case

Global life-sciences conglomerate

This (anonymous) bioscience multinational is leading the way thanks to automated phenotyping.

Home › Client cases › Global life-sciences conglomerate

Phenotyping is one of the activities of this international group operating worldwide in pharma, health, and biotech. Together with lab technicians, USoft developed an automated phenotyping solution.

The challenge

Lab technicians experiment with an almost infinite number of organic compounds. Each experiment contains photographs from multiple types of cameras. The lab technicians reported that they could no longer manage the huge volumes involved in high-speed phenotyping. Too much manual labor was involved in the process of experimentation, storage, and analysis, which caused them to fall behind. Moreover, organizing and searching through the data folders was problematic.

Too much manual work

Automatic phenotyping was not available. Phenotyping was manual work that took a lot of labour. This was a root cause for lagging image analysis.

The analysis of phenotyping images was lagging

More images were created than were analyzed. There were lots of images “on file” that had not been processed yet.
Images were not automatically checked for correctness. This means that faulty photos can be analyzed, corrupting experiment results.

Organising and searching the data folders was problematic

Files and images were hard to find. Long file names held metadata, and image organization was not automatic.
Before or after image processing, it wasn’t easy to find images for manual checks, reanalysis, or reviewing purposes.
Metadata was not stored centrally. This relates to the organization of the images in subfolders, preferably supported by a database structure that will make it easy to understand which image generates which data and where the image is.

Why this company chose USoft

Flexibility
One of the major advantages of USoft low-code is its flexibility. Lab technicians prefer to implement optimizations in processes independently, which is now possible.

More on low-code application development
High volume and high speed
Testing high volumes of substances in a structured, consistent, and traceable manner. Laborators face limited laboratory capacity and limited resources. They wanted to do more experiments in the same time.
Uniform practices
The business rules approach offered by the USoft platform is specifically suited for developing complex, data-intensive and rule-intensive applications. Because of using business rules, the main software solution is very agile, which makes future changes and additions easy to perform. If something changes, it is more transparent what is the effect on other pieces of the program.
More on business rules

The solution

USoft spoke extensively with the lab technicians responsible, and it quickly became apparent that this could be easily automated using USoft's low-code platform. The full project result is a state-of-the-art phenotyping system that will automatically organize- and analyze the correct images. This will enable researchers to perform analysis to determine the best strategy to reach their goals.

Data aquisition

Trial execution: preprocessing validation algorithm for image verification.
Preprocessing: image validation using validation algorithms for further processing.
Image organization: if the image is valuable, store the image and it’s metadata.

Data analysis

Preprocessing: image validation using validation algorithms for further processing.
Image organization: if the image is valuable, store the image and it’s metadata.

Reporting

Visualization/publishing API: make numerical result data available for further processing.

The results

Increased productivity
Analysis of +30.000 samples per day in a structured, consistent, and traceable manner. Decreased error margins by 80%.
Shortened time to market
Lab results are delivered 10x faster than before. That is why they are now a leader in product development.
Less manual work
Up to 95% repetitive operations and manual work have been removed from the process.