Related Experiment Video
Updated: Dec 31, 2025
14:06
Detection of Rare Genomic Variants from Pooled Sequencing Using SPLINTER
Published on: June 23, 2012
15.6K
Accurate detection of mosaic variants in sequencing data without matched controls
Yanmei Dou1, Minseok Kwon1, Rachel E Rodin2,3,4,5
1Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA.
Nature Biotechnology
|January 8, 2020
Summary
Detecting mosaic mutations during development is hard. MosaicForecast, a new machine-learning tool, accurately identifies these rare genetic variants, improving disease research.
Area of Science:
- Genomics
- Computational Biology
- Developmental Biology
Background:
- Detecting mosaic mutations, which occur in a small fraction of cells during development, is challenging due to the lack of matched controls and the presence of artifacts.
- Existing methods struggle with the low variant allele frequency and the need to distinguish somatic mutations from germline variants and technical noise.
Discussion:
- Mosaic mutations are increasingly recognized as important in normal development and disease pathogenesis.
- Accurate detection is crucial for understanding their functional impact and contribution to complex diseases.
Key Insights:
- MosaicForecast, a machine-learning method, uses read-based phasing and read-level features to detect mosaic single-nucleotide variants and indels.
- The method significantly improves specificity compared to existing algorithms.
- Validation using single-cell and targeted sequencing confirmed high accuracy for detected variants (80-90% for SNVs, 60-80% for indels).
Outlook:
- MosaicForecast offers a powerful new tool for the accurate detection of somatic mosaic mutations.
- This advancement will aid in elucidating the role of mosaic mutations in disease origins and development.
- Further applications in various biological contexts and disease research are anticipated.
