CHICAGO (GenomeWeb) – The CEO of a Taiwanese genomic informatics startup is in California this month in pursuit of funding and customers as the company prepares to enter the US market.
"For our product, for our software, you can only take full advantage of the capabilities if you have a large amount of sequence data to process," Atgenomix CEO and Cofounder Allen Chang said. The US is where many of the largest stores of data are — not to mention loads of venture capital.
Atgenomix, based in Taipei, hopes to raise $5 million in a Series A round, Chang said by phone while in Silicon Valley to meet with potential investors and users, including Stanford University.
"I've been visiting Silicon Valley frequently in the second half of this year," Chang said. He plans on flying to the East Coast next month to visit researchers at Harvard Medical School and other Boston-area institutions.
"Also, we have been looking for scientific advisors," he added. Two Stanford researchers are on board already.
Atgenomix provides "enterprise-ready parallel computing" for genomics to enable precision medicine, according to company marketing material.
"We want to provide a very solid, scalable infrastructure, a computational environment that can help the clinician-researcher as well as the clinical scientist process their genomics data," Chang explained.
"In the future, every patient's genome will be sequenced. When they go to see a doctor, they will have their sequencing data interpreted by a doctor and be provided a personalized medicine treatment," Chang continued. "The process and the speed of the processing as well as the quality of the data processing have become very important."
There has to be fast turnaround — less than a week — and the data has to be reliable. "If you don't change your computing infrastructure, it's not going to meet the requirement of one week to generate a report," Chang said.
Atgenomix claims to accelerate workflows 10-fold compared to traditional genomic analysis processes by making even standard desktop computers more efficient in processing genome data. "Accelerating precision medicine, right now I think it's the hottest topic in genomics, especially for whole-genome and whole-exome sequencing," Chang said.
"To run whole-genome sequencing, it requires several underlying technologies, not just the algorithms developed by the researchers," Chang said. "You need computational infrastructure, including the pair processing, as well as the data storage."
Researchers also need to be able to put everything together into a single, production-ready environment, Chang said. "Atgenomix provides enterprise-grade, enterprise-ready, proprietary software that puts everything together, not only accelerating the whole workflow, [but] we also provide integrated data storage and the deployment with all that information," he said.
Atgenomix offers both cloud and locally hosted options because some customers in Taiwan have expressed concerns about putting such sensitive patient-specific data into the cloud, according to Chang. "Bandwidth is still a bottleneck," Chang said. "The sequencing data is so huge, and it's going to be even more huge in the next generation of sequencers."
Illumina's latest product, the NovaSeq 6000 system, can produce as much as 6 terabytes of data over two days per WGS run. "The data is going to be bigger and bigger, and it requires a certain amount of bandwidth to upload data into the cloud," Chang noted.
Atgenomix started in April 2015. Chang and his cofounder, current Atgenomix chief science officer Chung Tsai Su, worked together in Taiwan at security software giant Trend Micro. Chang, a computer scientist who studied in Canada, was a senior global project manager during his 11-year stint with Trend Micro.
"We were dealing with a lot of big data. One day, we were having a conversation with a clinician and a researcher in Taiwan," Chang recalled. "They mentioned ... next-gen sequencing technology that can eventually transform precision medicine."
Chang and Chung saw an opportunity to contribute their expertise to "something which is maybe very meaningful," the CEO said.
Most clinicians and even clinical researchers don't have the necessary computing expertise or IT background, Chang noted. "They want to utilize the data for clinical practice, for medical practice, but the problem is, they just cannot handle such huge amounts of data," he said.
At Trend Micro, Chang and Chung used a lot of data-mining and machine-learning technology built in the Hadoop and Spark programming languages. "We could use the same technology to solve next-gen sequencing data-mining problems," Chang said.
"In a sense, the underlying technologies [for large-scale informatics] are all the same", he continued. "They're just different applications."
However, NGS may be the biggest of all challenges in big data.
"Everyone says it's just ATCG — four characters. But to deal with that, you require a lot of technologies to join together to help to process them very efficiently and correctly," Chang said. "We happen to have a very efficient algorithm that can speed up the sequencing data correlation and processing."
So far, Atgenomix has targeted pharmaceutical and diagnostics companies as well as research laboratories rather than practicing clinicians.
"Our current strategy ... is to focus on the reputable, large research labs that are doing a lot of genomic data processing," Chang said. "Not only to get some revenue, but most importantly, to have some reference users in the market."
Clinical practice will come later. "A lot of people are trying to figure out how they can effectively use sequencing data to diagnose disease and provide precision medicine," Chang said. "The market is still early and a lot of adoption is still not in production yet."