NEW YORK —Cloud-based data management firm Genomedia said recently that its services will be used in the third stage of SCRUM, the Cancer Genome Screening Project for Individualized Medicine in Japan.
The six-year-old company, based in Tokyo, has been supporting SCRUM-Japan since the project commenced in 2015. By participating in the project, Genomedia hopes to improve its services while winning over new customers, according to CEO Tomoyuki Yamada.
"We are developing our services through taking part in the project [and gaining] valuable feedback from users," he said.
Yamada cofounded Genomedia in 2013 to commercialize technology developed at the University of Tokyo and launched its Genomedia Front data management service two years later. In 2017, the firm raised ¥300 million ($2.5 million) in seed financing from JAFCO, a Japanese private equity firm that invests in life sciences companies. It currently employs about 30 people at its headquarters near the university in the city's Hongō district.
According to Yamada, Genomedia Front enables users to manage medical, laboratory, drug history, and genomic information, as well as personal data collected from patients, while supporting access to public and research databases and providing an information infrastructure for translational research.
"We started from building an integrated database of medical and genomics information, and we developed a system to annotate [the data] with public and private curated databases and to assist extracting knowledge from these databases," he said.
Users of Genomedia Front can look at data from numerous perspectives — according to certain gene mutations, for instance — and the service is capable of producing genetic, clinical trials, and drug information reports, he said. Genomedia Front also collects curated information for use in clinical trials, especially for Japanese and other Asian populations, and by using this data, the company can provide users with up-to-date treatment options, Yamada noted. This includes lists of medicines approved by Japan's Pharmaceuticals and Medical Devices Agency, the governmental body tasked with regulating drugs and medical devices.
Genomedia is also a certified service provider for Oxford Nanopore Technologies and maintains Illumina sequencing instruments in house. However, Yamada said that its cloud-based services remain the company's focus. "We collaborate with a limited number of customers using our wet lab," he said.
The introduction of Genomedia Front coincided with the launch of SCRUM-Japan, a collaborative project involving industry and academia to create a national patient registry database with quality-controlled genome and clinical information. Using the database, researchers aim to accelerate clinical trials of new drugs for treating various tumors.
Led by the country's National Cancer Center, the first two stages of SCRUM have included two projects — LC-SCRUM, focused on lung cancer, and GI-SCREEN, focused on gastrointestinal cancers. As part of these projects, patient samples have been interrogated using Thermo Fisher Scientific's Ion Torrent Oncomine Comprehensive Assay, which covers roughly 160 oncogenes.
The third stage of SCRUM-Japan, which began this year, includes two additional projects. LC-SCRUM-Asia will expand screening to Taiwan and other East Asian countries with the goal of including lung cancer patients who have rare genetic changes and codeveloping new therapies with medical institutions to treat them. MONSTAR-SCREEN, the second new project, involves creating a genetic diagnosis network for patients with various types of solid tumors, except lung cancer and gastrointestinal cancer. The MONSTAR-SCREEN project will include cancer-gene analysis following liquid biopsy, as well as gut microbiota analysis.
Katsuya Tsuchihara, chief of the division of translational informatics at Japan's National Cancer Center, said that the third phase of SCRUM-Japan will involve collecting data on an additional 2,500 lung cancers and 2,000 solid tumors in the next two years in collaboration with a dozen pharmaceutical and in vitro diagnostics companies.
"Our datasets will expand their variety from a one-time, tissue-based multigene panel to longitudinal liquid biopsy and microbiome analyses," he said. "The integration of these complex multiomics data with clinical information and visualizing them is our next challenge."
According to Tsuchihara, SCRUM-Japan has already collected more than 10,000 cases of lung and gastrointestinal cancer from about 250 hospitals in Japan in the SCRUM-Japan clinico-genomic database as part of its first two stages. He declined to provide a budget for the effort.
Data collected has included tissue-based, multigene panel testing data, clinical information obtained at registration, and yearly follow-up data, including treatment history and information about patient outcome. This data is shared among researchers in 17 collaborating pharmaceutical companies and 66 academic institutes. "Based on this information, patients have gotten into more than 50 clinical trials of gene-matched investigative new drugs," said Tsuchihara. In part as a result of SCRUM-Japan, six indications for five therapeutic drugs as well as six in vitro diagnostics have been approved in Japan to date, he noted.
"All these approvals were done based on the results of the clinical trials in which eligible patients were selected by the SCRUM-Japan genomic screening program," Tsuchihara said.
All data collected via SCRUM-Japan has been added to the cloud-based server provided via Genomedia. Tsuchihara said the investigators decided to use Genomedia Front in part because of its authentication and data security processes. The researchers, he noted, "required robustness for protecting patients' genomic data that were considered as 'special care-required personal information' by Japanese law."
Tsuchihara added that the database's filtering functions enable users to search for cases according to epidemiologic, clinical, and genomic information. Users that fulfill SCRUM-Japan's data sharing policy criteria can also download the trailable datasets and establish their own databases.
The work with Genomedia to refine and improve the database continues.
"We have been collaborating with the Genomedia team to customize and improve our database system for the last five years with satisfactory results," Tsuchihara said.
In a modularized way
To keep up with the demands of SCRUM-Japan, Genomedia has to constantly update its services, Yamada noted. To explain these updates, he provided a brief overview of how Genomedia Front operates. The first step involves aligning FASTQ reads to a reference genome and detecting mutations, insertion-deletions, and structural variants, including copy number variants. After the variants are called, the results are written to variant call format (VCF) files. These VCF files are annotated using numerous databases, including dbSNP, ClinVar, and others.
Once annotated, clinical information from electronic medical records is integrated with the genomic information, followed by other data, such as cancer type, age, gender, and medication history.
"Genomedia Front is developed in modularized way for these steps to respond quickly and flexibly to requests for various configuration changes in these pipelines," Yamada said. "It's possible to add or change items according to the application, and add plug-ins for original analysis, so customization can be realized as needed."
This will be useful in the third stage of SCRUM-Japan, which involves collecting genomic information from blood samples, or liquid biopsies, as opposed to solid tumors, the focus of the first two stages. Genomedia Front will update its service to enable the reporting of information obtained from minimal residual disease monitoring, cases with a high-risk of recurrence, early treatment response, and evaluation of drug resistance in real time.
The service now can provide users with an Integrated Cancer Genome Report that combines clinical, genomic, and other information with data on drugs corresponding to detected genetic mutation information, as well as ongoing clinical trial information.
"We aim to build a standard platform in the age of precision medicine through accumulating the achievements of the Integrated Cancer Genome Report," he said.
Yamada said that Genomedia is currently adding artificial intelligence functions to its services that will enable "squeezing new knowledge" from the database in a secure way. "We believe that in the future, our medical system will read genomic sequences much more than we are reading them now," he said. "We would like to be one of the leading companies to increase the value of medical data, including genomic sequences, using a secure, scalable, and robust cloud platform."
For now, Genomedia's market remains Japan, but the company hopes to expand abroad one day.
"Of course, we would like to go outside of Japan in the future with our services, because the cloud service could be accessed from all over the world," said Yamada. "But now we think the first step for [achieving] that is to collaborate with international companies who have already entered Japan or who would like to enter Japan."