A Supervised Success: Affymetrix Arrays Used to Predict Lymphoma Outcome
A new paper in the January issue of Nature Medicine presents research in which Affymetrix microarrays were used to profile gene expression in diffuse large B-cell lymphoma, and to characterize distinctive gene expression signatures in tumors that could be used to predict the outcome of chemotherapy.
The researchers, led by Margaret Shipp of Harvard’s Dana-Farber Cancer Institute, analyzed gene expression in diagnostic tumor specimens of patients that had received cyclophosphamide, adriamycin, vincristine and prednisone (CHOP)-based chemotherapy. They used a customized “lymphochip” array that included oligonucleotide probes for 6,817 genes specifically related to the germinal centers of lymphoid organs.
“The recent development of DNA microarrays provides an opportunity to take a genome-wide approach to predicting [diffuse large B-cell lymphoma] treatment outcome,” the researchers wrote.
The authors acknowledged that a group of investigators at Stanford’s Howard Hughes Medical Institute previously published a Nature paper on the same topic.
But while the Stanford group used a hierarchical (unsupervised) clustering algorithm to distinguish the gene expression profiles in the two different groups of diffuse large B-cell lymphoma, the researchers in this study also used a supervised learning prediction method. The method separated the patients according to gene expression patterns into two groups. In 58 of 77 patients whose tumors were profiled, the 5-year survival information was available. Those who had been classified by the algorithm as predicted to be cured had a 70 percent five-year survival rate compared to a mean survival rate of 54 percent, and those tumors that had been classified as fatal had a 12 percent survival rate.
This study not only adds to the body of research supporting the use of microarrays to predict outcome of diffuse large B-cell lymphoma, the researchers said, it also “illustrates the important difference between unsupervised (clustering) and supervised machine learning analytical approaches.”
With Cartesian Acquisition, Genomic Solutions Acquires New Stakeholders
When Genomic Solutions acquired Cartesian Technologies in late December, Schleicher & Schuell of Einbeck, Germany, in turn obtained a stake in Genomic Solutions.
Schleicher & Schuell formerly owned 29 percent of Cartesian’s outstanding common stock. Under the merger, this stock was exchanged for 6.25 percent of Genomic Solutions’ common stock, according to a filing with the US Securities and Exchange Commission.
Tom Tisone, the president of Irvine, Calif.-based Cartesian, also has acquired 5.69 percent or 1.77 million shares of Genomic Solutions’ stock, as part of the merger agreement. As of January 22, this would give Tisone a $4.12 million take in the company.
PAG Conference Attendees Show Animal Attraction to Microarrays
At the tenth annual Plant and Animal Genome Conference in San Diego last week, the halls were abuzz with talk of microarrays.
About 300 people showed up for a product demonstration in which Telechem/Arrayit.com of Sunnyvale, Calif. built a clean room equipped with PerkinElmer arraying technology, said Telechem/Arrayit business development vice president Paul Haje.
“More people are getting into microarrays,” Catherine Pettem, sales coordinator for biotechnology at Waterloo, Ontario-based Virtek Vision Corp, told BioArray News’ sister publication GenomeWeb. “People are becoming less dependent on core facilities. They no longer have to stand in line behind 10 researchers [to get microarrays].”
Robert Salazar, business development manager for bioscience products at Agilent Technologies, said attendees were quizzing him about custom microarray, and in particular those for Arabidopsis, which Agilent plans to introduce in the first half of 2002.
Array Express Database on Track for Debut at MGED Meeting in February
The Array Express database, the microarray database the European Bioinformatics Institute has been developing, is on track to debut at MGED 4, the Microarray Gene Expression Database working group’s annual conference, February 13 ¯ 15.
“We got the hardware delivery, the Oracle [system] has been installed and we are working on loading the datasets that we have at the moment,” said EBI microarray informatics team leader Elvis Brazma. “We are well on track to start taking data submissions by MGED 4.”