What is spatial clonal architecture?

The physical arrangement of a tumor’s genetically distinct subclones across its tissue, mapped so that each subclone’s territory, histological appearance, and surrounding microenvironment can be studied together rather than inferred from genetic data alone. A 2022 method combining whole-genome sequencing, in situ sequencing, and transcriptomics produced the first detailed whole-tumor-section maps of this kind.

How does tumor heterogeneity affect drug response?

Because different spatial regions of a tumor can carry different genetic alterations and different functional states, a treatment effective against the dominant, most accessible clone can leave a resistant, spatially distinct territory untouched. In glioblastoma, a spatially segregated region expressing a reactive-hypoxia program was found to carry more complex chromosomal alterations than the rest of the tumor, a combination classically associated with resistance.

Can spatial biology map tumor subclones?

Yes. Combining genome sequencing with in situ sequencing and transcriptomics has produced detailed, whole-tumor-section maps showing exactly where each genetic subclone sits, validated by microdissection. In breast ductal carcinoma in situ, this approach showed that distinct clones segregated within individual milk ducts rather than growing as an undifferentiated mass.

Tumor heterogeneity and spatial clonal architecture

Highlights in this article

Clonal evolution as a spatial process
Mapping subclones in tissue
Spatial patterns of resistance
Implications for combination therapy
Methods and caveats

Reconstructing spatial clonal architecture means treating a tumor’s genetic subclones not as an abstract list from a sequencing report but as physical territories with edges, neighborhoods and neighbors, each occupying real space within the tumor and interacting with whatever surrounds it. That physical framing turns out to matter enormously for drug development, because where a clone sits, not only which mutations define it, shapes how it behaves and how it responds to treatment.


Key takeaways A 2022 method combining genome sequencing, in situ sequencing, and transcriptomics generated the first whole-tumor-section maps of genetic subclone composition, validated by microdissection. In breast ductal carcinoma in situ, distinct clones expanded at a macroscopic scale but segregated within individual milk ducts, physical structures that shaped their growth. In glioblastoma, spatially segregated cells expressing a reactive-hypoxia program carry distinguishably more complex chromosomal alterations than their neighbors, a spatial pattern with direct relevance to treatment resistance. Early driver genetic events tend to span an entire tumor, while later changes remain regionally restricted, a timing-to-territory relationship with direct implications for which targets are defensible for combination therapy. Reconstructing clonal architecture spatially, rather than from multiregion sequencing alone, adds the histological and microenvironmental context that explains why a clone behaves the way it does.

Key takeaways

A 2022 method combining genome sequencing, in situ sequencing, and transcriptomics generated the first whole-tumor-section maps of genetic subclone composition, validated by microdissection.
In breast ductal carcinoma in situ, distinct clones expanded at a macroscopic scale but segregated within individual milk ducts, physical structures that shaped their growth.
In glioblastoma, spatially segregated cells expressing a reactive-hypoxia program carry distinguishably more complex chromosomal alterations than their neighbors, a spatial pattern with direct relevance to treatment resistance.
Early driver genetic events tend to span an entire tumor, while later changes remain regionally restricted, a timing-to-territory relationship with direct implications for which targets are defensible for combination therapy.
Reconstructing clonal architecture spatially, rather than from multiregion sequencing alone, adds the histological and microenvironmental context that explains why a clone behaves the way it does.

Clonal evolution as a spatial process

Genome sequencing of tumors routinely reveals mosaics of distinct subclones within the same mass, understood to arise through somatic evolution, essentially Darwinian selection acting on cells rather than organisms. What remained genuinely unclear until recently was the exact spatial growth pattern behind that evolution: not just which mutations a subclone carries, but where it grew, what physical structures shaped its expansion, and what it looked like histologically compared with its neighbors.

A landmark 2022 study in Nature addressed this directly by developing a workflow combining whole-genome sequencing, highly multiplexed base-specific in situ sequencing, and single-cell resolved transcriptomics, generating detailed quantitative maps of genetic subclone composition across entire tumor sections rather than scattered sampling points. Applied to 8 tissue sections from 2 multifocal primary breast cancers and validated by microdissection, the approach revealed intricate subclonal growth patterns that a conventional multiregion sequencing study, sampling only a handful of discrete points, would have been unable to resolve.

One finding from that work reframes what clonal evolution actually looks like physically. In a case of ductal carcinoma in situ, polyclonal neoplastic expansions occurred at the macroscopic scale, meaning multiple distinct clones were present across the visible lesion, but those clones segregated within microanatomical structures, specifically individual milk ducts. Evolution was not simply happening across the tumor as an undifferentiated mass; it was constrained and channeled by the tumor’s own physical architecture, with the ducts acting as microscopic corridors along which specific clones grew.

Mapping subclones in tissue

The practical value of mapping subclones spatially, rather than inferring their existence from bulk or multiregion sequencing alone, is that it adds exactly the context a genetic subclone list cannot supply on its own: what each clone looks like under a microscope, and what cells and structures surround it.

That same 2022 study demonstrated this directly. Across the progressive stages of ductal carcinoma in situ, invasive cancer, and lymph node metastasis, subclone territories were shown to exhibit distinct transcriptional and histological features and distinct cellular microenvironments. A subclone is not simply a genetic label attached to a group of cells; it is associated with a specific tissue appearance and a specific set of neighboring cell types, both of which are invisible to a sequencing report listing mutations alone.

Evidence type	What it establishes	What it cannot establish alone
Bulk or multiregion sequencing	Which genetic subclones are present, and their relative abundance in the sampled regions	The physical boundary of a subclone’s territory, or what surrounds it
Spatial genomics and transcriptomics	Where each subclone’s territory sits, its histological appearance, and its local microenvironment	Full whole-genome depth at every spatial position, since spatial platforms often trade some depth for spatial resolution
Combined approach	Genetic identity linked directly to physical territory, tissue structure, and surrounding cell composition	A single, static snapshot still cannot show how a territory changes over time without repeated or longitudinal sampling

Table 1. What spatial genomics adds to conventional sequencing-based clonal analysis, and its own remaining limitation. Combining approaches, not replacing one with the other, is the pattern that recurs throughout this evidence.

Spatial patterns of resistance

Clonal architecture becomes directly relevant to drug development when a specific spatial territory carries not just a distinct genetic identity but a distinct functional state associated with treatment resistance, and glioblastoma supplies a clear example of exactly that convergence.


When a spatial territory is both genomically complex and functionally distinct A spatial transcriptomic study of glioblastoma reconstructed clonal architecture using patient-specific hierarchical clustering of copy number aberrations. Cells expressing a reactive-hypoxia transcriptional program were found to be spatially segregated within the tumor, occupying a distinct territory rather than being distributed evenly, and this territory harbored distinguishably more complex chromosomal alterations compared with other spatial regions of the same tumor. That convergence, a specific transcriptional state, a specific physical location, and greater genomic complexity all co-occurring in the same territory, is mechanistically significant rather than coincidental. Hypoxic tumor regions are classically associated with treatment resistance across cancer types, and genomic complexity independently predicts more aggressive, harder-to-treat disease. Finding both properties concentrated in the same spatially defined territory means a treatment targeting the tumor’s dominant, more accessible clone may leave exactly this resistant, hypoxic pocket untouched, whatever its response elsewhere.

When a spatial territory is both genomically complex and functionally distinct

A spatial transcriptomic study of glioblastoma reconstructed clonal architecture using patient-specific hierarchical clustering of copy number aberrations. Cells expressing a reactive-hypoxia transcriptional program were found to be spatially segregated within the tumor, occupying a distinct territory rather than being distributed evenly, and this territory harbored distinguishably more complex chromosomal alterations compared with other spatial regions of the same tumor.

That convergence, a specific transcriptional state, a specific physical location, and greater genomic complexity all co-occurring in the same territory, is mechanistically significant rather than coincidental. Hypoxic tumor regions are classically associated with treatment resistance across cancer types, and genomic complexity independently predicts more aggressive, harder-to-treat disease. Finding both properties concentrated in the same spatially defined territory means a treatment targeting the tumor’s dominant, more accessible clone may leave exactly this resistant, hypoxic pocket untouched, whatever its response elsewhere.

Implications for combination therapy

A recent synthesis of glioblastoma spatial multi-omics work supplies a principle with direct, practical consequences for target selection in combination regimens: whole-tumor three-dimensional sampling has linked clonal evolution to physical territory, showing that early driver genetic events can span the entire lesion while later genetic changes remain regionally restricted.

That timing-to-territory relationship is a defensible, general argument for prioritizing early, truncal genetic alterations as combination therapy targets over late, subclonal ones, independent of any single cancer type. An early driver event, present since close to the tumor’s founding, is more likely to be found wherever a biopsy samples or wherever a therapy physically reaches, precisely because it was established before the tumor diversified into distinct territories. A late subclonal alteration, by contrast, may be confined to one region, and a single biopsy, or a therapy that does not distribute evenly throughout the tumor, can miss it entirely, leaving that region’s distinct biology unaddressed regardless of how well treatment performs elsewhere.

Three practical consequences follow for a program designing a combination regimen around clonal architecture evidence of this kind.

Prioritize targets confirmed as truncal or near-truncal. A target present across a tumor’s full spatial extent, rather than confined to one territory, is a more defensible single-agent or backbone target precisely because treatment reaching only part of the tumor still reaches it.
Treat regionally restricted alterations as combination candidates, not backbone targets. A late, spatially confined subclone is better addressed by a second agent specifically chosen to cover that territory’s distinct biology, rather than assumed to respond to whatever addresses the tumor’s dominant clone.
Treat resistant spatial niches as a design constraint from the outset. The hypoxic, genomically complex glioblastoma territory described above is exactly the kind of region a combination regimen should be designed to reach deliberately, rather than discovered as an unexplained source of relapse after the fact.

Methods and caveats

Two limitations deserve explicit statement before this evidence is applied to a real program, since both affect how confidently a spatial clonal architecture finding can be generalized.

Sample sizes in foundational spatial genomics work remain small. The 2022 breast cancer study validating this whole approach examined 8 tissue sections from 2 patients. That is sufficient to demonstrate that spatially organized clonal architecture exists and can be mapped, but insufficient on its own to establish how frequently the specific patterns observed, such as duct-constrained clonal segregation, recur across a broader patient population.
A tissue section remains a static snapshot. Even a perfect spatial map of clonal architecture describes one moment in a tumor’s ongoing evolution. Understanding how a specific territory changes under treatment pressure over time requires either repeated sampling or, where clinically feasible, paired pre- and post-treatment specimens, neither of which a single spatial genomics snapshot alone provides.

The broader disease-heterogeneity foundation this spoke builds on, including evidence spanning multiple non-oncology diseases and the specific tissue-sampling caution it raises, is developed in Understanding Disease Heterogeneity With Spatial Biology. For the broader tumor microenvironment argument surrounding clonal architecture, see Spatial Biology in Oncology: Decoding the Tumor Microenvironment, and for where this sits within the full spatial biology pipeline, Spatial Biology in Drug Discovery: From target discovery to translational medicine.

This article was produced under Drug Discovery News’s editorial policies.