NEW YORK (GenomeWeb) – Newly minted bioinformatics firm Signifikance recently launched Signifikance Insights, its software-as-a-service platform for analyzing and interpreting cancer genomic variants from whole-genome sequence in clinical contexts.
The company launched the tool early last month and currently has four unnamed customers, mostly from the clinical laboratory market, who are using the system in their cancer clinical workflows, Iker Huerga, the company's CEO, told GenomeWeb. Although its initial focus is the cancer testing market, the company is also looking to expand its business into the prenatal screening and testing market. It is currently testing the solution for that purpose in a pilot project with one of its current customers, Huerga said. It expects to launch a product for the prenatal market this year but is still mulling whether or not that will be a separate offering or marketed as part of its existing cancer platform.
Signifikance Insights is available primarily under a software-as-a-service model but customers with security concerns who prefer to keep their datasets behind their firewalls have the option to install the company's solution on a local server. The local system connects to Signifikance's main servers so that these customers have access to the daily updates that the company makes.
Under the SaaS option, customers are charged $99 per sample. Once generated, the report gets stored in the company's servers and the user can come back at any time and regenerate it free of charge. Patient reports are updated as new information about treatments or trials becomes available. Pricing for the local installation of the software is flexible and depends on the customers' needs and requirements, the firm said.
Signifikance Insights accepts variant call files as input and uses an internally developed machine-learning algorithm to identify clinically actionable variants that are specific to the patient in question. At the end of the analysis, oncologists receive a clinical report that contains a list of the clinically significant genetic variants for the particular patients, information on possible targeted therapies and treatment strategies, and options for clinical trials that would be most relevant for the patient based on their molecular profile and primary diagnosis. The information included in the report is backed with supporting information from scientific literature as well as multiple variant databases and repositories.
The company is currently in the process of securing the intellectual property rights to the algorithm, so it's not discussing the mechanics of the approach in much detail. But the algorithm is basically trained to look at lists of variants, identify pathogenic variants, and make decisions about whether the variants in question are relevant to the cancer based on information available in public repositories such as COSMIC and ClinVar among others.
It essentially replicates and automates what has historically been a manual curation process, Huerga explained. So instead of having researchers themselves sort through the information contained in the variant databases manually to identify which mutations are known to be pathogenic for a specific type of cancer, "what we have done is teach a machine to replicate that process and to be able to go through the evidence and to make a decision," he said.
The system also includes a knowledgebase that integrates information from more than 60 data sources and pulls in new information from these resources as they are updated. In addition to COSMIC and ClinVar, the list of databases that the company uses includes the National Center for Biotechnology Information and the Cancer Genome Atlas.
On a daily basis, "we run our custom machine-learning algorithm on top of all these data [and] extract new relationships, such as genotype-phenotype associations, so we are able to identify new variants that are known to be associated with a specific disease," Huerga said. The algorithm is also "able to distinguish between a variant that is likely to be pathogenic for a particular phenotype [and] a variant that is known to be benign or that is known to be a risk factor." As new associations are found, the company adds them to the knowledgebase and makes the information available to oncologists.
Signifikance officially opened its doors about a year ago with support from Blueprint Health, a startup accelerator program in NYC, and internal funding. Signifikance currently has four employees and hopes to add three more people to its workforce by the end of Q3 this year. It is looking specifically for people to work on business development and on product development related to its planned prenatal testing and screening offering, Huerga said.
As it moves into the next-generation sequencing-based testing market, the company will compete for customers with existing companies, such as Cypher Genomics and Personalis, which offer clinical cancer analysis and interpretation solutions, as well as firms like Omicia, which offers a clinical platform for reporting on causative variants from NGS gene panels, exomes, and whole genomes.
However, Signifikance believes that its solution can compete favorably in the marketplace by helping customers cut down on combined costs of sequencing and analyzing patient samples. Some firms, Huerga noted, offer their analysis services as part of a larger package that covers both the sequencing and analysis, and that could be a costly option for some potential customers, he said. As an alternative, a clinical lab could work directly with a provider that could sequence the sample for a fraction of the cost that the full-service company might charge and then send the resulting vcf to Signifikance and pay only an extra $99 to obtain a clinically actionable report, he said.
The company also expects some competition from firms that offer traditional NGS analysis services more generally. But since its offering is tailored to the clinical oncology market, it expects to be able to woo oncologists who are looking for tools that help them make decisions about patient care and treatment rather than solutions designed with the academic research market in mind, Huerga said.