NEW YORK (GenomeWeb) – Toronto, Canada-based biotechnology startup Ranomics is commercializing functional data on potentially pathogenic variants that have not been seen in clinical testing but could be implicated in disease development.
Ranomics aspires to offer functional data on all possible variants that could occur in human disease genes. For now, though, it has its sights set on characterizing mutations in roughly 20 to 25 common hereditary cancer genes over the next few years, starting with BRCA1, CEO Cathy Tie told GenomeWeb. This week, the company launched Ranomics Ring-Domain, a database of internally conducted high-throughput functional studies of 2,000 potential mutations in BRCA1, one of the most commonly mutated genes in breast and ovarian cancers. These studies assess the possible effects of these mutations on patients in model organisms such as yeast.
The company's pitch to clinical testing laboratories and omics data providers is that functional assays provide strong experimental information on these variants that will help them more quickly categorize and classify them if and when they do show up in genetic test results. "A lot of genetic tests are inconclusive because there is not enough information about a lot of the genetic variants that are analyzed in these tests," Tie said. Databases like ClinVar provide information on mutations that have been seen in clinical cases and are classified based on clinical criteria but "there are a lot of rare or unique mutations that [researchers] have never seen before so they don't know what it does to the human body" she said. These databases also do not provide functional information on these variants, she noted
Ranomics estimates that some 60 percent of hereditary cancer genetic tests find new mutations with no available genotype or phenotype information. In 2014, researchers from Memorial Sloan Kettering, the Mayo Clinic, and elsewhere launched the Participants in the Prospective Registry of MultiPlex Testing initiative in part to capture information on variants of unknown significance found in next-generation sequencing panel tests.
Last year, Quest Diagnostics and France's Inserm launched BRCA Share to provide a forum for clinical labs to share BRCA data and help speed up the classification of rarely seen mutations found in the gene. At that time, GenomeWeb reported that Inserm's Universal Mutation Database, the planned home for the lab's data, contained data on around 7,200 BRCA1 variants, about 1,400 of which were categorized as VUS.
On the other side of the spectrum, Myriad Genetics claims to have the lowest VUS rate, compared to its competitors, with 0.6 percent of tests identifying a VUS in BRCA1, according to 2013 numbers. Myriad has about 46,000 variants in its proprietary database, gleaned from testing more than 2 million patients so far — about 17,000 of those come from BRCA1 and2. But "even if we sequence everyone on the planet, there is always going to be this long tail of more rare and unique mutations that people have because everyone has different DNA," Tie argued.
Functional studies can help researchers evaluate the effects of these mutations and are a valuable complement to predictive algorithms and familial testing. "There is always going to be a rare or unique mutation that has never been seen before, [and] that's where we come in," she said.
Ranomics uses a PCR-based method to generate all of the possible amino acid changes that could occur at each position in a given region of the gene of interest, Tie explained. In the case of BRCA1, the company focused on the Ring domain, a region of the gene commonly mutated in breast and ovarian cancer cases. The exact mechanism that the company uses to generate the mutations is proprietary, but Ranomics CSO Leo Wan explained via email that the researchers use oligonucleotide primers to swap amino acids at each position in the gene.
The next step is to insert the mutant gene into a model organism and then analyze the resulting phenotype to determine whether the mutation is harmful or not. For the BRCA1 mutations, Ranomics researchers used yeast cells as the model organism and measured the cells' growth to assess gain or loss of function. Harmful genetic mutations cause a much stronger growth phenotype in yeast compared to benign mutations, which do not lead to as much growth, Tie explained.
Ranomics' researchers quantified the mutants' pathogenicity by comparing the growth curves from yeast assays to growth curves from known benign and pathogenic mutations. They scored the variants on a scale of zero to one, depending on how the assay and clinical growth curves aligned. A zero score meant that the mutation was likely benign, while a score of one meant the variant was likely pathogenic.
"We can provide data on every possible genetic mutation instead of just the clinical ones," she said. "We provide this unique set of data that's generated from functional studies on these rare or unique variants that people have never seen before. We are able to generate them in the lab and test them in model organisms instead of waiting for it to be seen in a real human and then collecting that clinical data, which is what the other databases do."
Tie and Wan co-founded Ranomics — a portmanteau of 'rare' and 'genomics' — in 2014. They came up with the idea for the company as students at the University of Toronto. Tie, then an undergraduate majoring in bioinformatics, was studying human BRCA1 mutation expression using yeast assays when she met Wan, who is working on a doctorate in the department of molecular genetics at the university.
Tie has since left school to focus on growing Ranomics' business. She is a recipient of the Thiel fellowship, a project of Paypal co-founder Peter Thiel that offers students 22 or younger a $100,000 grant to stop or skip college for two years and start new businesses. The company has raised an additional $100,000 from IndieBio, an accelerator program run by SOSV, a venture capital and investment management firm. It has also raised an undisclosed amount of seed money from angel investors and venture capital firms.
Tie and Wan first became interested in the genetic testing space around the time consumer genetic testing firm 23andMe ran afoul of the US Food and Drug Administration for failing to address questions about the clinical and analytical validity of its direct-to-consumer service. The testing firm's troubles with the FDA got Ranomics' co-founders thinking about challenges in the genetic testing industry, one of which is dealing with VUS, Tie told GenomeWeb.
So far, at least two companies have signed partnership agreements with Ranomics. The company reached out to French bioinformatics firm Interactive Biosoftware and New York City-based SolveBio, both of whom have agreed to offer access to its BRCA1 data to customers of their respective platforms. Interactive Biosoftware markets the Alamut software suite of tools for analyzing, annotating, and interpreting genetic variants, while SolveBio provides curated datasets from public and proprietary resources, including genes, proteins, metabolites, drug compounds, clinical trials, and variant-disease relationships.
Ranomics' novel approach to the genetic testing market was attractive to Solve Bio CEO and co-founder Mark Kaganovich. "It's one of the only models that I'm familiar with that does these kind of comprehensive experiments on variants of unknown significance in genes that people are interested in," he told GenomeWeb. It is a model he believes will gain ground as testing gets more tailored. "Obviously, clinical observations are probably the strongest evidence that one can bring to bear for a variant classification, but barring access to those kinds of clinical observations ... experimental evidence is useful and better than nothing."
For Interactive Biosoftware's founder and CEO André Blavier, high-throughput functional assays are a valuable tool for generating useful data for more accurate clinical diagnostics. "For years, [there has been] research done on in silico predictions of the functional effects of variations, but this moves at a very slow pace and does not provide very reliable information," he told GenomeWeb. "Functional assays [offer] much stronger evidence for clinical use, so that makes everything really interesting here." Since Ranomics is yet to publish any data, Blavier said his firm will monitor clients' response to the Ranomics Ring Domain database and gauge their interest in receiving more of the company's data in the coming months. "We [want] to see what clinicians will say, and if they find that the functional data correlates with their own findings," he said.
Ranomics does plan to publish data from its studies sometime in the future, Tie said. The company is currently collaborating with a number of clinical labs in the US, among them Veritas Genetics. Ranomics recently concluded a pilot study with Veritas where it performed functional studies of BRCA1 variants provided by the clinical lab. The study aimed to demonstrate the "strength and reliability" of functional studies in variant classification efforts, according to Ranomics. Veritas is now using Ranomics' functional data in its variant classification process, becoming one of the first clinical labs to do so.
Meanwhile, the company is generating more functional data for BRCA1 variants. Tie said the company is now using a human cell line assay to run more functional studies on the gene. It is also developing another assay for functional studies on TP53, another commonly mutated hereditary cancer gene, she said.
Clinical testing laboratories can purchase Ranomics' variant data directly from the company's website, where the company charges $100 per variant. For that price, customers receive a variant report that includes a functional score and a suggested pathogenic or benign classification for the variant in question. Ranomics also offers a bulk-subscription option, where pricing ranges from $1,500 to $2,500 per year, depending on the size of the lab and its needs. Ranomics also offers an application programming interface for software companies who want to include its data into their offerings.
This article has been updated to correct the spelling of Peter Thiel's name.