Related Experiment Video
Updated: Mar 19, 2026
Detecting Somatic Genetic Alterations in Tumor Specimens by Exon Capture and Massively Parallel Sequencing
Published on: October 18, 2013
Genome-wide classification of tumor-derived reads from bulk long-read sequencing
Toby M Baker1, Nedas Matulionis1, Cassidy Andrasz1,2
1Division of Hematology Oncology, Department of Medicine, David Geffen School of Medicine, University of California Los Angeles.
Abstract:
DNA extracted from tissue samples typically derives from a complex mixture of cell types. Without single cell analysis, it has been generally impossible to determine the cell type of origin for most molecules. One clear example of this is in the complex milieu of a human neoplasm. Here, we develop ROCIT (https://github.com/tobybaker/rocit), a transformer-based model to classify the tumor or non-tumor origin of individual reads from bulk tumor samples sequenced with long-read whole genome sequencing. Using somatic mutations to derive training data, ROCIT uses read-level methylation patterns to accurately classify reads from anywhere in the genome without requiring the adjacent normal tissue or the explicit identification of tumor differentially methylated regions. We apply ROCIT to a cohort of prostate and ovarian tumors and demonstrate high classification accuracy across the entire genome. We then demonstrate the potential of ROCIT predictions to improve somatic variant calling. ROCIT represents a major step forward in the analysis of bulk tumors with long-reads, enabling the accurate and sensitive identification of reads with specific cell types of origin genome-wide.
Insights
Researchers developed ROCIT, a transformer model for classifying DNA origin in bulk tumors. This method uses methylation patterns from long-read sequencing to distinguish tumor from non-tumor cells genome-wide, improving variant calling.
Area of Science:
- Genomics
- Bioinformatics
- Cancer Research
Background:
- Bulk tumor DNA is a mixture of cell types, making cell-type origin determination difficult without single-cell analysis.
- Human neoplasms present a complex cellular environment where distinguishing tumor from non-tumor DNA is challenging.
Purpose of the Study:
- To develop a computational model (ROCIT) for classifying the origin of individual DNA reads from bulk tumor samples.
- To enable accurate genome-wide classification of reads originating from tumor versus non-tumor cells using long-read sequencing data.
Main Methods:
- Developed ROCIT, a transformer-based model utilizing read-level methylation patterns for classification.
- Trained ROCIT using somatic mutations identified in bulk tumor samples.
- Applied ROCIT to long-read whole genome sequencing data from prostate and ovarian tumors.
Main Results:
- ROCIT achieved high classification accuracy for read origin (tumor vs. non-tumor) across the entire genome.
- The model accurately classifies reads without needing adjacent normal tissue or pre-identified differentially methylated regions.
- ROCIT predictions demonstrated potential to enhance somatic variant calling accuracy.
Conclusions:
- ROCIT is a significant advancement for analyzing bulk tumors using long-read sequencing.
- The model facilitates accurate and sensitive identification of cell types of origin for DNA reads genome-wide.
- This approach offers a powerful new tool for understanding tumor heterogeneity and improving genomic analyses.
