A validated lineage-derived somatic truth data set enables benchmarking in cancer genome analysis

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Existing cancer benchmark datasets rely on germline variants or synthetic methods, which are inadequate for accurate somatic mutation analysis.
• Current validation methods for somatic mutations are often expensive and lack comprehensive coverage across the whole genome.

Purpose of the Study:

• To introduce the Lineage derived Somatic Truth (LinST) dataset, providing a reliable resource for studying somatic mutations.
• To establish a benchmark dataset for human sequencing data that accurately reflects true somatic variation.

Main Methods:

• Development of the LinST dataset using the HT115 colon cancer cell-line.
• Validation of somatic mutations through a known cell lineage, encompassing thousands of mutations.
• Establishment of a high-confidence region of 2.7 gigabases per sample for comprehensive analysis.

Main Results:

• The LinST dataset contains a large collection of validated short somatic mutations.
• The dataset leverages a known cell lineage for high-confidence mutation identification.
• A significant portion of the genome (2.7 Gb) is covered by high-confidence variant data per sample.

Conclusions:

• The LinST dataset provides a superior resource for benchmarking somatic mutation callers and cancer genomics studies.
• This validated dataset addresses the critical need for accurate whole-genome somatic variation data.
• LinST facilitates advancements in understanding cancer genetics and developing targeted therapies.