CHICAGO – Next week, GeneXplain will release Genome Enhancer, a modified and simplified version of its flagship multi-omics bioinformatics platform, to support clinicians in their pursuit of precision medicine.
The Wolfenbüttel, Germany-based company was responding to market demand from biologists and other molecular laboratory professionals without deep bioinformatics backgrounds, according to GeneXplain Cofounder and CEO Edgar Wingender.
"We noticed that our software platform, which is a very complex toolbox, was too complex for the end users, the life scientists who want to play around with their own data rather than to approach an expert nearby," Wingender said. "This toolbox is good for experts, but way too complex for the end users, for the biologists, [and] for the bench workers who are generating the data and would like to do something with these data themselves," Wingender said.
He described the new Genome Enhancer as a "one-button solution" for such clinicians and lab specialists.
"The life scientist just enters the data, presses a button, then gets a full-fledged report about his data and the [reasons] that may have caused the phenomenon he was studying." Wingender said.
A sample gene expression analysis report provided by Wingender provided detail on "promising druggable targets" for treating melanoma, based on gene regulators that Genome Enhancer identified. The report, written much like a journal article, detailed the transcriptomic and proteomic analysis and featured numerous charts and graphs to illustrate findings and probabilities.
GeneXplain is positioning Genome Enhancer for precision medicine in clinical settings, not just research laboratories.
Cofounder and Chief Scientific Officer Alexander Kel said that unlike many other tools for understanding genomic mutations, Genome Enhancer can interpret variations in gene regulators, not just mutations in the actual genes being studied. This capability comes from a machine learning engine in the software.
"With this approach, we can find new drug targets for a patient," Kel said. "If we apply it to precision medicine, we can actually even deal with individual data, data coming from one patient, and propose what kind of treatment could be beneficial for this patient based on their gene expression data."
Genome Enhancer is part of the flagship GeneXplain platform and draws on all the same resources.
"It includes everything starting from variable data and ends up writing a full report with interpretation of the results," Kel said. "It's a complete and complex pipeline. Although it is technically done using the GeneXplain platform, we have it as a separate product."
Kel said that he expects clinicians and some researchers to start with the new Genome Enhancer because it is simple to use. More knowledgeable users are able to move over to the core GeneXplain platform for more detail or to modify pipelines, Kel added.
The full GeneXplain platform features a complete workflow management system, as well as hundreds of pipelines for various bioinformatics tasks.
"You can connect many different programs, tools, and scripts with each other. This is now one optimized workflow, which is … important for medical applications in the future of precision medicine," Wingender said.
Genome Enhancer includes a proprietary process that GeneXplain calls "upstream analysis." It focuses on identifying gene regulators.
"You have a set of differentially expressed genes. You compare, say, a liver cell to a tumor cell and you see that in the tumor cell that there are 300 genes are differentially upregulated," Wingender said.
"Other tools may take these genes and map them to Gene Ontology categories, for instance. We take the genes and analyze their upstream sequences," he explained.
"When we have identified these regulators, we go into the regulatory network of the cell" in search of master regulators, Wingender continued. "In nearly all cases studied so far, we identified upstream the pathways that regulate the activity of the transcription factors controlling these gene convergence points — genes or molecules that regulate exactly this group of genes that were identified as differentially regulated."
In a paper published in April in BMC Bioinformatics, a team including Wingender and Kel, plus researchers from the Institute of Chemical Biology and Fundamental Medicine in Novosibirsk, Russia, and the Bellvitge Biomedical Research Institute in Barcelona, Spain, described how they apply upstream analysis to detect new biomarkers for early-onset colorectal cancer.
In that study, the researchers studied RNA sequences and DNA methylation data on tumor samples from 300 patients from Denmark at various stages of colorectal cancer progression. They then ran the data through an early iteration of Genome Enhancer.
The machine learning built into the software was able to identify a minimum set of six biomarkers — ALCA, ENO1, MYC, PDX1, TCF7, and ZNF43 — which, when tested on a relatively small cohort of 12 tumor samples and 12 controls, showed sensitivity and specificity levels of at least 92 percent. This set, the researchers wrote, offered the "best cancer detection potential."
They did find some limitations of the technology, however. They admitted to having "rather simple methods" of finding transcription factor binding sites in DNA sequences, which resulted in a modest 36 percent improvement in identification of such sites.
This process requires the Transfac database, a manually curated collection of eukaryotic transcription factors and related binding information. GeneXplain's bioinformatics and machine learning looks for patterns in the database to analyze regulatory sequences.
"There is really three decades of accumulated knowledge in this application. That is something we are a little bit proud of," Wingender said. GeneXplain leans heavily on technology Wingender helped develop at the University of Göttingen.
Indeed, prior to starting GeneXplain in 2010, Wingender was director of the university's Institute of Bioinformatics, which spun out a molecular database company called Biobase.
Qiagen acquired Biobase in 2014. Biobase created and Qiagen still maintains the Transcription Factor Binding Sites database, which contains information on eukaryotic transcription factors and miRNAs.
In 2016, Qiagen granted GeneXplain a worldwide, exclusive, and perpetual license to maintain, develop, and commercialize Transfac, Transpath, and Proteome, all former Biobase databases.
"We decided to split [GeneXplain] off because Biobase was completely in the database business." Wingender explained. GeneXplain is a software-oriented company.
Biobase also had a hand in the development of BRENDA, a collection of enzyme functional data, among other resources. Continued BRENDA development maintenance is in the hands of the Institute of Biochemistry and Bioinformatics at the Technical University of Braunschweig, Germany, but GeneXplain technology supports processing of BRENDA pipelines.
All of those databases and more are now fully integrated with GeneXplain's software platforms, according to Wingender.
GeneXplain had some grant funding from the German government and the European Commission in its early days, but has long been self-sustaining, debt-free, and has never needed outside investors, according to Wingender.