PLASQ: a generalized linear model-based procedure to determine allelic dosage in cancer cells from SNP array data

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Human cancers often arise from acquired mutations, including copy number polymorphisms.
• Copy number polymorphisms involve deviations from the normal diploid state of autosomal chromosomes.

Purpose of the Study:

• To develop a novel procedure for quantifying parental chromosome copy number contributions in cancer cells.
• To leverage single-nucleotide polymorphism (SNP) microarray data for precise copy number analysis.

Main Methods:

• Developed probe-level allele-specific quantitation (PLASQ) using a generalized linear model.
• Employed a novel probe classification and an expectation-maximization algorithm for model fitting.
• Applied the method to SNP microarray data from various cell types, including lung cancer.

Main Results:

• Demonstrated strong model fit across diverse cell types.
• Achieved highly accurate genotyping in normal diploid samples.
• Enabled generalized genotyping in cancer samples, identifying events like amplified SNPs.
• Validated findings through independent computational and experimental methods.

Conclusions:

• PLASQ is an effective method for analyzing parental chromosome copy number contributions in cancer.
• The developed procedure offers high accuracy in both normal and cancer samples.
• An R software package for PLASQ is publicly available for research use.