CHICAGO (GenomeWeb) – After more than five years in operation, French biomedical analytics startup OmicX recently made a key business decision.
"At the end of last year, we started to monetize the platform," said CIO Philippe Berenguel, who joined OmicX in 2018. The platform, a multilevel search engine for omics-related knowledge, now operates on a freemium model.
"You can find most of our pages right from Google and land on it and access the data, but you don't have all the navigational and additional data aspects like that," Berenguel explained. "If you want to access that, you have to pay a fee."
The company, which is headquartered in Le-Petit-Quevilly, France, near Rouen, now offers 24-hour passes and annual subscriptions, based on the observation that users tend to return sporadically.
"A user won't be coming back every week, but he might be coming back every quarter or every two months. When he does, he tends to use the platform very intensely to find an answer to a specific question," Berenguel said.
OmicX Founder and CEO Arnaud Desfeux said that the company's mission is to provide better insights to biologists and biomedical researchers, on account of the fact that the volume of data has exploded over the last decade. "The ability of biologists to analyze this data is not as good as the ability to generate the data," said Desfeux, who holds a PhD in biology.
As for the platform itself, OmicX applies manual curation and artificial intelligence to search the internet for medical literature and index freely available bioinformatics tools, in what the company calls a "dynamic analytics ecosystem," based on mining "collective intelligence" to uncover scientific insights.
"For a few years now, we've collected as many bioinformatic tools as we could find in the literature and we classified them in a database. Through these manually curated bioinformatic tools, we were able to recreate how biologists and bioinfomaticians do their analysis," Desfeux said, likening the platform to Google Maps for navigating large biomedical datasets by finding pathways described in published literature.
"[With a navigational app], when you want to go from Point A to Point B, you are going to have different checkpoints you are going to have to go through," he said. OmicX does similar by showing users what tools and pathways are available for answering their research-related questions.
"What we give biologists is a huge gain of time because instead of having to find all the publications themselves to do the analysis, they now can do it in one click on our website," Desfeux said.
"We are able to create new ways of analyzing the data. When we gather all that information from various publications, we also can sort it out based on how often we have seen it in the literature. We can create unique protocols or pipelines of rules to analyze the data that the biologist has," Desfeux said.
For institutions performing complex literature searches, time savings could be counted in months, according to Aurélien Quillet, a molecular biologist who leads R&D and AI development at the company.
"Let's say you are looking for a mutation analysis in breast cancer," Quillet said in an online demonstration of the technology for GenomeWeb. "Then you get [a list of] the various resources that we give out to the biologists." This includes types of software that others have used to study the topic or that OmicX has classified through manual curation as being suitable for such an application.
OmicX AI has automatically analyzed protocols described in scientific literature to find those that could work with such a study.
"We also annotated the publication with different sets of diseases, genes, or whatever molecule has been studied. That way, through our search engine, we are able to understand that this publication was used for studying mutations in breast cancer and it can give you ideas on how to analyze your own data," Quillet said.
This technology, company officials said, is far superior to PubMed and consumer search engines. OmicX is not unique in this space, though; others offering genomic search engines include Genomenon, IQvia's Linguamatics unit, and Saphetor.
"You are removing a lot of noise in the background coming from search engines like Google," according to Berenguel. He noted that general search engines rank results by popularity, meaning that papers not widely cited could be buried deep in the results. "With an automated extraction tool like ours, if it's mentioned even one time in the literature, we will find it," he said.
With Google, "You would have to go through all the lists of publications that the search engine is going to give you, read them, analyze them, get all the tools and possibilities that are being used, then try it," Quillet said. This creates a tedious cycle of trial and error.
OmicX tries to cut out that extra work, not only by looking for outliers, but by listing probabilities when looking for, say a pipeline for mutation-line research into breast cancer, helping researchers get an idea of the scientific consensus, according to Quillet.
"In research, you always have to prove that the method you are using is really the best you can find so far for what you are doing. We offer this proof for you because we give you all the publications associated with your research and the protocols," he said. "We link to you the references, so if you have to prove something, you can base your choice on the literature, not just on what we tell you."
Chief Operating Officer Marion Denorme, a neuroscientist who previously was a researcher at INSERM (French National Institute of Health and Medical Research), said that OmicX to date has indexed and listed on its website more than 100,000 pages of original content about software, protocols, and datasets available to researchers.
OmicX has about 1 million annual users and upwards of 25,000 subscribers, about a 50-50 split between computer scientists and biologists, according to Desfeux. The company gets about 2,000 new subscribers per month, he said.
Most users are academic researchers, according to Berenguel. About a third are in Europe, a third from the Americas, and the rest from Asia. The US and China represent the two largest country-specific user bases, he said.
OmicX has raised €4.5 million ($5.1 million) in grant and venture funding since 2013. Sources include a technology fund for the Normandy region of France, the European Union's Horizon 2020 research and innovation program, and some unspecified private investors in Switzerland.
While the current OmicX product is focused on search, the company also is working on an analytics service for the data it indexes and on developing an application programming interface so third parties could develop their own apps based on OmicX's data, Berenguel said.
At the moment, OmicX is working mostly with open-access papers, so the company also will be exploring how to index the content in literature protected by publisher paywalls, according to Quillet.