ABACUS: an Automated Algorithm for Determining Genotypes from Quantitative Probe Hybridization Data

An automated statistical method has been developed to determine genotypes of individual DNA samples from quantitative hybridization assays to DNA probes. This method was validated using Affymetrix Variation Detection Arrays (VDAs). The method provides a quality score (i.e. confidence score) to individual genotype calls, allowing investigators to focus their attention on sites that give accurate and repeatable information. Description (Set) The invention is an automated statistical framework for determining genotypes. It has been applied to Affymetrix VDAs that quantitatively measure oligonucleotide hybridization between target and probe sequences. The VDAs probe for all possible DNA variants at a given site. The technique can be applied to any technology that assays multiple probes in a quantitative manner, including virtually any quantitative fluorescence or mass spectrometry technology. ABACUS has already been applied to Affymetrix VDAs and HuSNP arrays (GeneChip). Obvious variations to ABACUS include relaxation of the Gaussian distribution assumption, maximum likelihood estimation of background intensity scaling factors, and minor coding modifications to improve execution speed. The computer code implementation of ABACUS is written in ANSI C, and can be compiled and run on any modem computer. Proposed Use (Set) ABACUS provides an objective framework for deciding whether or not the differences in signal intensity are sufficiently large to make a reliable inference about the genotype of a sample while also bringing likely VDA errors to the attention of the investigator. To this end, it provides a quality score for all genotype calls that is a measure of the statistical reliability of that call.

Inventor(s): Cutler, David

Type of Offer: Licensing

« More Engineering Patents

Share on      

CrowdSell Your Patent