Reconstructing tumor evolutionary histories and clone trees in polynomial-time with SubMARine

Area of Science:

• Computational Biology
• Cancer Genomics
• Evolutionary Biology

Background:

• Tumors comprise diverse cancer cell subpopulations with distinct genetic profiles.
• Understanding cancer evolution is crucial for effective treatment strategies.
• Current clone tree methods struggle with uncertainty and scalability.

Purpose of the Study:

• To develop a scalable method for reconstructing cancer evolutionary histories.
• To address the limitations of existing clone tree reconstruction techniques.
• To represent and summarize uncertainty in cancer evolution inference.

Main Methods:

• Formalized partially-defined clone trees to represent sets of evolutionary histories.
• Introduced the Maximally-Constrained Ancestral Reconstruction (MAR) to summarize equally fitting clone trees.
• Developed SubMARine, a polynomial-time algorithm to approximate the MAR, handling copy number aberrations and noisy frequency estimates.

Main Results:

• SubMARine efficiently reconstructs cancer clone trees, scaling polynomially with the number of subclones.
• The algorithm accurately recovers known evolutionary histories and identifies errors in existing reconstructions.
• Demonstrated high performance on simulated data, lung cancer, and prostate cancer datasets, running in under 70 seconds.

Conclusions:

• SubMARine offers a scalable and robust approach to inferring cancer evolutionary dynamics.
• The method effectively captures uncertainty in clone tree reconstruction.
• Provides a valuable tool for cancer research with open-source implementation available.