Three well-established statistics companies are getting ready to stake their claims in the genomics market with customized products for gene expression analysis and statistical genetics.
Just last month, SAS released SAS Genetics for the analysis of genetic markers, including linkage disequilibrium and haplotype reconstruction, and has already attracted several customers, according to Russ Wolfinger, director of SAS’s genomics group.
Wolfinger has led this R&D group of ten for a year and a half, and has been fostering collaborations with various academic institutions in North Carolina’s Research Triangle to learn about the statistical needs of biologists and to co-publish studies. Another product, SAS Microarray Solutions, a combination of statistical methods, data warehousing and management, and visualization tools, will be released in the fall, and similar solutions are in the works for proteomics, he said.
Even though there is no shortage of gene expression analysis software, Wolfinger claimed that SAS has an edge with its high-end statistics and ability to process very large datasets in high-throughput mode. “There is definitely a trend now in the community to move toward a more rigorous statistical type of software analysis,” he said. “We have got a framework which I think is pretty powerful, and a lot of the statistical methods are generic.”
S-Plus Adds Arrays
“The whole area of microarray analysis is on a fast track toward statisticians,” agreed Michael O’Connell, who heads the pharmaceutical group of data analysis company Insightful. His company is planning to launch a microarray toolkit based on its flagship statistics product S-Plus in September. It will contain tools for annotating and normalizing microarray data, differential expression analysis, data visualization, and other applications, he said.
The need for such a toolkit, he said, arose over the last two years as non-clinical statisticians in pharmaceutical and biotechnology companies were increasingly tasked with analyzing microarray data. These traditional customers of S-Plus have in fact frequently written their own software for various applications, O’Connell said, and Insightful’s microarray toolkit is “an immediate response to our customers’ needs.”
Moreover, the company has several early-phase NIH grants to develop “more sophisticated tools” for microarray analysis. It just received a two-year $750,000 SBIR grant to develop statistical genetics software for the study of pedigree data based on linkage and linkage disequilibrium analysis. The aim is to develop a comprehensive package of tools, since “the software that’s available is mostly academic software, and it’s very hard to use,” said Antje Hoering, principal scientist for Insightful’s genetics project.
SAS and Insightful are not the only companies customizing their software for gene expression analysis. Within the next few months, statistics company SPSS is planning to launch a microarray application template — a collection of “streams” or analyses covering several steps of microarray data analysis — for its Clementine data-mining workbench.
According to Cathy DeSesa, senior marketing analyst for SPSS Science, Clementine is more flexible and comprehensive than existing gene expression analysis software. Owing to an external module interface, users can add other components, such as modeling algorithms, visualization tools, or data access routines, and they can combine models. To showcase Clementine’s capabilities, SPSS has been collaborating with Duke University’s Bioinformatics Shared Resource since the spring, where Duke researchers receive access to Clementine and share the results of their analyses and methods with SPSS. The main purpose of this partnership is “to be able to publicize the work that’s been done [using Clementine] and not just have it come from a vendor,” said DeSesa.
The space for gene expression analysis software is quickly growing even more crowded, but statistics players are confident that demand for their software will grow as well: “There is plenty of room for all of us to have success in that,” said Wolfinger. Given the large expenses for instruments and reagents, “a little bit more for the software, I think, would be well worth the investment,” he said.