Tumors are not static, uniform masses of identical cells. Instead, they behave like rapidly evolving ecosystems. As a tumor grows, individual cells continuously acquire random somatic mutations. Because these mutations are passed down to all subsequent daughter cells, they act as permanent, heritable molecular barcodes.
By reading these barcodes across different cells, regions, or time points, researchers can perform phylogenetic clonal reconstruction. This process turns a static genomic snapshot into a historical movie, mapping out the lineage tree of a cancer to reveal exactly how it began, diversified, and spread.
The Anatomy of a Lineage Tree: Trunk vs. Branch
The core intuition behind clonal reconstruction relies on separating the "trunk" of an evolutionary tree from its "branches."
Imagine a mutation that occurs in the very first malignant cell. As that cell divides, every single subsequent cancer cell inherits that alteration. This is a truncal mutation, representing the clonal origin of the disease. Conversely, a mutation that occurs later in a single subpopulation of cells will only be present in a fraction of the tumor. This is a branch mutation, representing subclonal diversification.
[First Cancer Cell] <-- Truncal Mutation (Shared by all)
/ \
/ \
[Clone A] [Clone B] <-- Branch Mutations (Subclonal/Private)
To mathematically sort the trunk from the branches, algorithms analyze the Variant Allele Frequency (VAF)—the proportion of sequencing reads that carry a specific mutation. As Constance H. Li demonstrates in a study on tumor hypoxia (Divergent mutational processes distinguish hypoxic and normoxic tumours), tracking VAF distributions allows researchers to map the architecture of genetic changes over time, revealing how selective pressures shape the tumor.
Identifying this hierarchy is clinically vital. As Jing Ren et al. (Copy number variations in esophageal squamous cell carcinoma: Emerging cancer drivers and biomarkers (Review)) note, therapeutic strategies must prioritize truncal alterations. Targeting a branch mutation only eliminates a subset of cells, leaving other subclones free to drive a recurrence; targeting a truncal mutation ensures the therapy strikes the entire tumor population.
Capturing the Evolutionary Snapshot
To build an accurate tree, researchers must capture the genetic diversity within a patient. This is achieved through several distinct sampling strategies:
Multi-Region and Spatial Sampling
By physically separating distinct areas of a tumor, researchers can compare mutations across space to infer history. Gregor Warsow et al. (Genomic features of renal cell carcinoma with venous tumor thrombus) utilized multi-region sequencing to trace whether dangerous anatomical extensions, like venous thrombi, arose from existing primary subclones or represented independent evolutionary events.
To achieve even finer resolution, Hamzeh Kayhanian et al. (Homopolymer switches mediate adaptive mutability in mismatch repair-deficient colorectal cancer) coupled laser capture microdissection with whole-exome sequencing, reconstructing high-resolution trees from localized subclonal patches. Taking this to the organ scale, Andrew Erickson et al. (Spatially resolved clonal copy number alterations in benign and malignant tissue) treated spatial transcriptomic profiles as molecular barcodes, mapping lineages across physical tissue space to show that genetically aberrant clones often expand into histologically "benign" tissue long before visible structural changes occur.
Single-Cell Sequencing
Bulk sequencing averages out the genetic signals of thousands of cells. Single-cell sequencing bypasses this limitation by observing exactly which mutations co-occur within individual cells. Ravindra Majeti et al. (Mutation order in acute myeloid leukemia identifies uncommon patterns of evolution and illuminates phenotypic heterogeneity) leveraged single-cell DNA sequencing to determine the precise chronological order of genetic events in leukemia, proving that the sequence of mutation acquisition shapes the clinical personality of the disease.
Furthermore, Yifan Zhao et al. (High-resolution detection of copy number alterations in single cells with HiScanner) demonstrated that integrating single-nucleotide variants with fine-scale copy number alterations (CNAs) at the single-cell level provides an ultra-high-resolution roadmap of tumor progression.
Liquid Biopsies
Tumors continuously shed DNA fragments into the bloodstream. Muhammed Murtaza et al. (Multifocal clonal evolution characterized using circulating tumour DNA in a case of metastatic breast cancer) demonstrated that circulating tumor DNA (ctDNA) accurately mirrors the evolutionary hierarchy of metastatic cancer. Truncal mutations common to all tumor sites appear at higher plasma concentrations than site-specific branch mutations, enabling non-invasive, real-time tracking of systemic lineage evolution.
The Mathematics of Tree Building
Once mutational data is gathered, computational frameworks convert these frequencies into hierarchical trees.
One foundational approach is Dollo parsimony. This model assumes that complex genetic alterations (such as specific structural variants) are highly unlikely to be independently gained multiple times across different lineages, but they can be lost. Ikram Ullah et al. (Evolutionary history of metastatic breast cancer reveals minimal seeding from axillary lymph nodes) applied Dollo parsimony to map the relationships between primary tumors and metastatic sites, revealing whether metastases spread linearly from one secondary site to another or in parallel directly from the primary hub.
However, tumor evolution is noisy, and ancestral populations rarely disappear completely; they often persist alongside their descendants. To capture this biological reality, Ke Yuan et al. (BitPhylogeny: a probabilistic framework for reconstructing intra-tumor phylogenies) introduced BitPhylogeny, a Bayesian framework that allows cancer clones to be represented as internal nodes within the tree rather than just terminal leaves.
Standard Tree (Leaves only): BitPhylogeny (Internal Nodes Persist):
[Ancestor] [Clone A (Ancestor)]
/ \ / \
/ \ / \
[Clone A] [Clone B] [Clone B] [Clone A* (Evolved)]
Similarly, Khalil Ouardini et al. (Reconstructing unobserved cellular states from paired single-cell lineage tracing and transcriptomics data) developed TreeVAE, a probabilistic model that uses lineage constraints to accurately infer the unobserved historical transcriptomic states of ancestral cells.
To ensure these reconstructed trees reflect true biology rather than sequencing artifacts, validation tools are required. Alice Antonello et al. (Computational validation of clonal and subclonal copy number alterations from bulk tumor sequencing using CNAqc) created CNAqc, a quality control framework that checks if subclonal copy number calls are mathematically consistent with somatic mutation frequencies under standard evolutionary models.
Resolving Clinical Mysteries
Lineage mapping answers fundamental questions about how diseases manifest and resist treatment in individual patients.
Metastasis vs. Independent Primaries
When a patient presents with multiple tumors, determining whether they are independent primary growths or metastases from a single source drastically alters treatment. Ren‐Chin Wu et al. (Elucidating the pathogenesis of synchronous and metachronous tumors in a woman with endometrioid carcinomas using a whole-exome sequencing approach) used phylogenetic reconstruction to prove that a tumor appearing eight years after an initial diagnosis was not a late recurrence, but a second primary tumor sharing a distant common ancestor that had seeded benign tissue years prior.
Similarly, Anqiang Wang et al. (Molecular characteristics of synchronous multiple gastric cancer) utilized subclonal population analysis to classify multiple synchronous gastric tumors into monoclonal (metastatic) or polyclonal (independent genetic) pathways.
Developmental Mosaicism
Sometimes, shared mutations do not imply metastasis or inherited germline risk, but rather an early embryonic event. Risa Burr et al. (Developmental mosaicism underlying EGFR-mutant lung cancer presenting with multiple primary tumors) quantified shared versus private passenger mutations to show that multiple primary lung tumors in non-smokers can arise from a single mutation occurring early in embryonic development, creating a localized genetic predisposition.
Phenotypic Plasticity
Lineage trees can also prove that tumors change their physical identity to escape treatment. Emily Robitschek et al. (916 Acquired resistance to immune checkpoint blockade by phenotypic plasticity of melanoma) used phylogenetic modeling to confirm that a muscle-like rhabdomyosarcoma lesion and a melanoma lesion in the same patient shared a common clonal origin. This proved that the tumor evaded immunotherapy via non-genetic shape-shifting rather than acquiring new driver mutations, demonstrating the power of lineage tracing to uncover hidden mechanisms of therapeutic resistance.
Key Foundational Work
Reconstruction of Cell Lineage Trees in Mice (2008) — contributed 1 concept elements to this direction.
TrAp: a tree approach for fingerprinting subclonal tumor composition (2013) — contributed 1 concept elements to this direction.