CNARA: reliability assessment for genomic copy number profiles.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• DNA copy number profiles from microarrays and sequencing can contain wave artefacts and large derivative log ratio spread (DLRS), compromising data reliability.
• These artefacts, often introduced during sample preparation or due to poor DNA quality, are not fully removable by current preprocessing methods.
• Misidentification of copy number alterations/variations (CNA/CNV) due to these issues necessitates methods for reliable data assessment.

Purpose of the Study:

• To develop and validate a computational method for distinguishing reliable genomic copy number profiles from those with significant artefacts or poor quality.
• To provide a tool that aids in the exclusion of unreliable samples or adaptation of downstream analysis strategies in genomic screening experiments.

Main Methods:

• A novel method was developed using four key features to summarize copy number profiles for reliability assessment.
• A classifier was trained on a dataset of 1522 copy number profiles from diverse microarray platforms.
• The method is designed to be platform-independent for microarrays and adaptable for sequencing platforms generating piecewise constant copy number signals.

Main Results:

• The developed method effectively distinguishes reliable from unreliable genomic copy number profiles.
• The classifier demonstrates the ability to predict profile reliability irrespective of the microarray platform used.
• The approach is adaptable for sequencing data, enhancing its broad applicability in genomic studies.

Conclusions:

• A new method, CNARA, has been developed for assessing the reliability of genomic copy number profiling data.
• This method should be applied alongside existing quality control procedures to improve genomic data mining.
• CNARA can enhance the reliable functional attribution of copy number aberrations, particularly in cancer research.