SuperFreq

SuperFreq performs integrated detection of somatic single nucleotide variants (SNVs), short indels, copy number alterations (CNAs), and clonal tracking from exome sequencing data to characterize tumor clonal evolution across multiple samples.

Key Features:

• R package implementation: Implemented as an R package for analysis of cancer exomes.
• Integrated analysis pipeline: Identifies somatic SNVs, short indels, and CNAs from exome sequencing data and integrates these results for clonal analysis.
• Clonal tracking without matched normals: Operates without requiring paired tumor-normal samples by utilizing unrelated control samples.
• Cross-checking of variant calls: Cross-checks variant calls across multiple samples from the same patient to distinguish somatic from germline variants and to resolve overlapping CNA calls.
• Performance on TCGA exomes: Demonstrated high recall and precision on 304 cancer-normal exome samples across 33 cancer types from The Cancer Genome Atlas (TCGA), reliably identifying clones with a cellular fraction of at least 50% and retaining robustness when matched normals are unavailable.
• Validation via in silico mixing: Shown in simulation studies using in silico mixing of cancer and normal samples to assign mutations to the correct clone with high accuracy.
• Broad applicability: Applicable beyond exomes to other capture libraries and suitable for analyses such as tracking diagnosis and relapse samples in leukemia.

Scientific Applications:

• Clonal evolution analysis: Reconstruction and tracking of distinct cellular clones across multiple tumor samples to characterize temporal and spatial tumor evolution.
• Tumor heterogeneity and resistance studies: Investigation of intra-tumor heterogeneity and mechanisms underlying treatment resistance through mutation and CNA assignment to clones.
• Studies without matched normals: Enabling genomic analyses of cancers where matched normal samples are unavailable by using unrelated controls.
• Diagnosis–relapse comparisons: Comparative analysis of diagnosis and relapse samples in diseases such as leukemia to track clonal dynamics.

Methodology:

Implemented as an R package that identifies somatic SNVs, short indels, and CNAs from exome sequencing data, integrates these calls for clonal tracking, uses unrelated control samples instead of matched normals, cross-checks variant calls across multiple samples to distinguish somatic from germline and resolve overlapping CNAs, and was validated on 304 TCGA exomes and by in silico mixing simulations.