NEW YORK (GenomeWeb) – Belgian software developer UgenTec is now offering software that uses artificial intelligence-based machine learning for end-point PCR analysis, such as SNP calling.
UgenTec is currently targeting the software, called FastTyper, at the ag-bio sector but said that it may also be useful in clinical labs. The firm has also supported a recent evaluation that found its flagship PCR workflow product, FastFinder, can save labs up to 90 percent of analysis time.
UgenTec CEO Steven Verhoeven said that the company developed FastTyper after it received requests from players in the ag-bio sector who perform food testing and animal testing using extremely high-throughput PCR tests. The software is meant to meet the needs of SNP genotyping in high-throughput settings — on the order of millions of tests per year — helping companies to analyze those test results efficiently.
UgenTec has already signed customers that are also co-developing the product, so that it is "tuned to the exact market need," Verhoeven said. One collaborator evaluating FastTyper is VHLGenetics, a firm that offers genetic tests for agriculture and farm applications such as phenotyping or detecting inherited diseases in its labs in Belgium, Germany, and The Netherlands.
For SNP genotyping specifically, the firm's artificial intelligence software is trained on billions of historic datapoints so that it can accurately process whether a specific result is a mutant type or wild type, for example. According to Thomas Beuls, a marketing manager at UgenTec, incorporating historic results sets the software apart from competing technologies, which tend to use simple thresholding and only reference data from within a single run.
FastTyper also deploys on-line learning, such that it starts off trained on multiple variants of data, but over time, as user input is given to it, the algorithm will tune itself, and essentially "get smarter," Beuls said.
UgenTec also sees FastTyper being used for SNP genotyping in a clinical setting eventually. "The biggest demand at this point in time is coming from the ag-bio sector, but we are definitely making it part of our suite of products from UgenTec so that it can be used by different types of companies and different types of industries,"
Founded in 2014 as a spinoff of Leuven University, UgenTec used funds raised in 2016 to further develop its flagship product, FastFinder PCR analysis software. The firm also raised $9.3 million in a Series A funding round earlier this year. Overall, it has raised $13 million over three different rounds, and is currently active in more than 35 countries with approximately 200 labs using its software to interpret results, Verhoeven said.
The problem, as Verhoeven described it, is that many molecular diagnostics laboratories and diagnostics companies rely on scripts, in-house built software, and Excel macros to interpret results of molecular tests.
"We saw an opportunity here to use techniques like artificial intelligence to provide intelligent software solutions to automate and standardize the PCR workflow," Verhoeven said.
The flagship FastFinder software helps labs "to be more efficient in their PCR workflow" by automating the interpretation of raw data. It is particularly useful for multiplex PCR tests, or syndromic panels, which can yield reams of data that require highly trained lab technicians to interpret.
"We work together with diagnostic companies and molecular laboratories to seamlessly integrate and connect the different components of the PCR workflow, support interpretation of their test results, and in the end, decrease the time they need to perform this interpretation, to perform the workflow, and increase reliability and decrease cost," Verhoeven said.
FastFinder uses AI to analyze real-time PCR data. "We trained it on over a billion data points and hundreds of assays so that it can take the result of any PCR device and any type of assay and truly automate the interpretation," Verhoeven said. The large datasets have come from customers using devices from companies such as Agilent, Bio-Rad, Roche, and Abbott, he said.
The software specifically trains a model to automatically combine information across multiple curves and types of fluorescent signals and channels, only highlighting those samples that need manual investigation.
Verhoeven said it is difficult to give an exact cost for the product at this time the value will be based "on time savings and the increased reliability that customers get from it," and in the end will be a fraction of a customer's operating cost.
The software's main competitor, according to Beuls, is likely the status quo of labs entrenched in less efficient protocols, as well as Microsoft Excel-based approaches.
He noted that Fast Track, now a Siemens subsidiary, presented data last week at the European Society for Clinical Virology meeting showing the company could shave off 90 percent of the analysis time on its multiplex tests using the FastFinder software.
The firms had previously launched a molecular testing platform called Fast Track Cycler that is provided with FastFinder software at the European Congress of Clinical Microbiology and Infectious Diseases in April.
UgenTec has also recently expanded its vision to include middleware and related software.
"Historically, our solution was mainly focused on interpretation support and using artificial intelligence to capture the expertise of a lab technician and implement that in automated software to support that technician," Verhoeven said. Now the firm is "expanding into middleware, and really connecting our solution to the different instruments and different software applications that the molecular laboratory uses," he said.
Along these lines, UgenTec recently created a consortium called the Molecular Automation Network. In partnership with Hamilton Robotics, the network will facilitate the exchange of business and technical information between the manufacturers of automated liquid handling workstations, real-time PCR devices, software, and diagnostics. Specifically, UgenTec and Hamilton will work together to develop new business opportunities, exchange technical expertise, and integrate their existing software and hardware systems, as previously reported.