Skip to main content
Premium Trial:

Request an Annual Quote

Data Mining Saffron Spices Up Portfolio with Life Sciences

Premium

If you can make sense of the government, you can do just about anything. At least that’s what the folks at Saffron Technology are bargaining on. With a Research Triangle Park, NC-based technical team founded in 1999 and comprised mostly of former IBMers, Saffron has been proving its data-mining technology with projects at NIH and US intelligence agencies for a couple of years, and is now looking to expand into the bioinformatics sector.

“The requirements in the two fields are very similar,” says CTO Toufic Boubez. “They both have a lot of data — a lot of it is unstructured, and even when it is structured, it’s [in] very different structures.”

Boubez thought about bioinformatics when he read an NIH paper on cancer cell work. “I was a little bit surprised,” he recalls. “Ninety percent or more of that paper was devoted to numerical computation.” To see if the data problems really were similar to what Saffron had already been targeting, Boubez and his teammates ran part of the NIH analysis through the company’s engine and wound up with the same answer in significantly less time.

The key to Saffron’s products, a machine-learning engine and a network that builds on it, is the lack of parameters. “It’s based on associative memory techniques,” Boubez says, “and takes away all the problems with traditional data mining.” Instead of building models and attempting to fit the data into them, Saffron’s technology watches data — and then when a user asks a question, the engine “builds a model on the fly,” according to Boubez. With the larger network product, anything — a gene, a person, a document — is assigned to a piece of software whose job is to watch and learn about its relationships and connections.

Saffron’s technology runs on any computer that’ll run Java, Boubez says. “All our projects are built on straight Pentium machines.” The 15-person company has one customer so far in life sciences, an instrument manufacturer in the Northeast. And in its other area, the company just got a sole-source label after a year-long due diligence project with the NSA, which means Saffron’s technology is so unique that it won’t have to go through the bidding process to work for the government.

“We’ve proven our technology in a different domain that’s just as hard, if not harder, [than bioinformatics],” Boubez says. “We’ve got the legs, we’ve got the proofs.”

— Meredith Salisbury

 

The Scan

Interfering With Invasive Mussels

The Chicago Tribune reports that researchers are studying whether RNA interference- or CRISPR-based approaches can combat invasive freshwater mussels.

Participation Analysis

A new study finds that women tend to participate less at scientific meetings but that some changes can lead to increased involvement, the Guardian reports.

Right Whales' Decline

A research study plans to use genetic analysis to gain insight into population decline among North American right whales, according to CBC.

Science Papers Tie Rare Mutations to Short Stature, Immunodeficiency; Present Single-Cell Transcriptomics Map

In Science this week: pair of mutations in one gene uncovered in brothers with short stature and immunodeficiency, and more.