CONIPHER: a computational framework for scalable phylogenetic reconstruction with error correction.

Area of Science:

• Genomics
• Computational Biology
• Cancer Research

Background:

• Intratumor heterogeneity drives cancer evolution.
• Accurate inference of tumor subclonal architecture and evolutionary history from DNA sequencing is challenging due to artifacts and errors.
• Existing methods struggle with copy number alterations and mutation losses, impacting phylogenetic reconstruction.

Purpose of the Study:

• To present CONIPHER, a computational framework for accurate inference of tumor subclonal structure and phylogenetic relationships.
• To address challenges in tumor evolutionary history reconstruction from multisample DNA sequencing data.
• To provide a robust tool for analyzing complex cancer genomes.

Main Methods:

• Development of CONIPHER (COrrecting Noise In PHylogenetic Evaluation and Reconstruction) computational framework.
• Accounting for copy number alterations and mutation errors in phylogenetic analysis.
• Application to multisample tumor sequencing data, including whole-exome sequencing and primary-metastatic cases.

Main Results:

• CONIPHER accurately infers subclonal structure and phylogenetic relationships.
• The framework was successfully applied to reconstruct tumor phylogeny from the TRACERx421 dataset.
• CONIPHER outperforms similar methods on simulated data and scales to large datasets.

Conclusions:

• CONIPHER provides an accurate and robust method for reconstructing tumor evolutionary histories.
• The framework enables automated phylogenetic analysis for large-scale sequencing datasets across different technologies.
• CONIPHER is accessible to users with basic bioinformatics and scripting knowledge.