Bioinformatics startup Tinybio has developed a novel software program that enables scientists to run complicated information analyses just by asking questions in plain English, similar to speaking to a educated assistant. The corporate’s CEO and founder, Sasha Dagayev, says the motivation behind Tinybio is to make information evaluation sooner and extra accessible.
“We are very much a tool for scientists,” Dagayev defined in an interview. “The basic premise is the ability to kind of chat your samples and interrogate your data using natural language.”
At its core, Tinybio’s software program makes use of giant language fashions and pure language processing to translate a person’s requests into scripts that run widespread bioinformatics instruments like R, Samtools, and FastQC behind the scenes on highly effective compute situations. Nonetheless, in contrast to conventional command-line interfaces, Tinybio presents a clear chat interface paying homage to ChatGPT.
“It’s accessible to everybody. You can just sign up with a Google account,” stated Dagayev. “We’re hooked up to a machine on Google’s infrastructure that actually is doing the analysis for you.”
Whereas Dagayev acknowledges there are present “no code” instruments, he argues they’re restricted in scope. “No-code tools exist, but they box you into a world where you can only do what the no-code tool can do. Whereas in reality, a lot of the stuff that people work on is outside of that by the nature of their job.”
Tinybio has already attracted 500 month-to-month customers in its first 12 months, 25 of whom pay for premium entry. A giant a part of the preliminary goal market, in keeping with Dagayev, is college students and early-career researchers nonetheless studying to navigate bioinformatics workflows.
“It’s easier to get a younger person to change the way they’re working,” he stated. College students are sometimes challenged with utilizing a device like R studio with little to no coaching. Whereas business scientists may need entry to for-profit instruments with help contracts, college students are sometimes left to determine issues out on their very own.
Past making information evaluation extra accessible by pure language interfaces, Dagayev hopes Tinybio’s know-how can present insights into how science will get accomplished by analyzing the requests customers make.
“We know things that nobody else in the world knows,” he stated. “For example, Thursdays is our number one day, and we think that’s because most lab meetings happen on Fridays.”
As Tinybio features extra traction, the corporate plans to increase its capabilities to allow evaluation for brand new information sorts like spatial transcriptomics. Dagayev can also be exploring potential partnerships with corporations like Oxford Nanopore to bundle Tinybio’s software program with genomic sequencing applied sciences.
