New Mathematical Model Shows Promise for Cancer Genomics

Armen Hareyan's picture


Some researchers work "by the numbers" to identify the genetic hallmarks of cancer. Now, a powerful mathematical tool introduced by Broad Institute scientists will facilitate this task, providing a clearer picture of how DNA runs amok in tumors.


The genetic distortions that lurk within the cells of a tumor form the driving force behind malignancy. These changes, which involve either gains or losses of DNA, perturb the usual number of gene copies in a cell and can involve either of the paired chromosomes. Scientists are trying to trace the chromosomal origins of such modifications to pinpoint informative genes, forming the basis for new therapeutic targets and possible genetic predictors for cancer diagnosis.

Researchers led by Matthew Meyerson, an associate member of the Broad Institute and associate professor at Harvard Medical School/Dana-Farber Cancer Institute, developed an algorithm that interprets the data from single nucleotide polymorphism (SNP) arrays, a collection of short oligonucleotides ("oligos") used to tally SNP patterns in human DNA. Probe-level allele-specific quantitation (PLASQ), described in the November issue of PLoS Computational Biology, allows scientists to approximate DNA copy number at sites throughout the genome and to assign the proportion furnished by each parental chromosome.

"In cancer, genome modifications often affect only one of the two paired chromosomes, the one inherited from the father or the one contributed by the mother," said Meyerson. "PLASQ allows us to localize these changes to the culprit chromosome, which will help guide us to the most significant genes and gene mutations in the disease."

In SNP arrays, oligos are parsed into probe sets and each set, comprised of 40 probes, is tailored to detect a single SNP in the human genome. These probe sets consist of one probe that perfectly matches both the target SNP and its surrounding DNA, as well as several related probes, each harboring single letter mismatches. To compute DNA copy number, PLASQ exploits the mathematical link between a probe's intensity