NEW YORK (GenomeWeb) – Smpl Bio is preparing to launch a software solution that will help researchers select the optimal candidates for their targeted sequencing, qPCR panels, and other kinds of biomarker platforms.
The company, which formed a year ago, is commercializing machine-learning algorithms for biomarker selection that were developed by researchers at the University of Connecticut and which are described in a BMC Bioinformatics paper published last November. These tools have been combined together into an easy to use pipeline that does not require programmatic expertise to run and will hopefully help simplify experimental assay design for researchers developing targeted platforms, James Lindsay, the company's co-founder and chief architect, told GenomeWeb.
According to Smpl Bio, the pipeline efficiently selects predictive biomarkers for disease identification, drug response prediction, and sample classification, among other tasks from high-dimensional datasets and includes tools for data quality control and data visualization. The platform also shows researchers how predictive the suggested biomarkers are for their particular research question; lets them compare the costs and benefits of using the biomarkers in their panels, and allows them to tweak their selections depending on what best fits their experiments and budgets, according to the company. The system also provides researchers with information on the impact of choosing fewer or more biomarkers in their assays, Lindsay told GenomeWeb.
Smpl Bio intends to offer the system under a software-as-a-service model after it has validated it. The pipeline is currently being validated in a series of projects being run by researchers at the University of Connecticut. Once those are completed, Smpl Bio plans to offer its tools under a business-to-business model, Lindsay said, meaning that it will license its pipeline to companies that offer targeted biomarker platforms such as Fluidigm, Illumina, and Life Technologies.
These companies offer software that helps researchers with experimental design tasks, such as primer design, as well as selecting the right consumables. But they don't provide solutions for biomarker selection, leaving that task to the biologists themselves, which isn't always the best solution, Lindsay noted. Smpl Bio's intent is to get its system integrated into these firms' existing consumables ordering pipeline where it can be offered as an option to customers of their targeted platforms.
Smpl Bio is actively speaking with a number of unnamed industry partners at present, Lindsay said, and won't launch its pipeline until it has firm agreements in place. Pricing for licenses to the system is still being worked out, he said adding that partner companies can choose how and if they want to pass those licensing costs on to their customers.
When its product comes to market, Smpl Bio expects to compete with existing open-source solutions that are used for prioritizing biomarkers, such as the Broad Institute's GenePattern, Lindsay said, as well as general statistical tools that can be customized for biomarker selection. However, these tools generally aren't easy to use and require that users have programming expertise, he said. In addition, while general statistical tools offered by firms such as SAS and Google can be adapted for biomarker selection, they aren't customized for use in the biological arena and offer little to no support for the data in this domain, he added.
On the commercial bioinformatics front, companies such as Partek, Golden Helix, and JMP Genomics also offer statistical packages for biomarker prioritization. However, Smpl Bio's platform provides a more comprehensive service that covers quality control, advanced statistical visualizations, data filtering, and validation of the selected biomarkers, Lindsay said.
Smpl Bio's offices are located just outside of Hartford, Conn. The company currently has three full-time employees and is actively recruiting candidates for a sales position and a bioinformatics vacancy.