Detection of chromosomal abnormalities using high resolution arrays in clinical cancer research

In clinical cancer research, high throughput genomic technologies are increasingly used to identify copy number aberrations. However, the admixture of tumor and stromal cells and the inherent karyotypic heterogeneity of most of the solid tumor samples make this task highly challenging. Here, we propose a robust two-step strategy to detect copy number aberrations in such a context. A spatial mixture model is first used to fit the preprocessed data. Then, a calling algorithm is applied to classify the genomic segments in three biologically meaningful states (copy loss, copy gain and modal copy). The results of a simulation study show the good properties of the proposed procedure with complex patterns of genomic aberrations. The interest of the proposed procedure in clinical cancer research is then illustrated by the analysis of real lung adenocarcinoma samples.

Graphical abstractGenomic profile of a lung adenocarcinoma sample.Figure optionsDownload full-size imageDownload as PowerPoint slideHighlights►High throughput genomic technologies are used to identify copy number aberrations. ► We propose a robust two-step strategy for this task. ► Results of a simulation study show the good properties of the proposed method. ► Lung adenocarcinoma samples are used to illustrate the interest of the procedure.