For decades, genomic analysis has relied on bulk sequencing, a method that aggregates genetic material from thousands or millions of cells into a single measurement. While powerful, this approach creates an "averaging" effect that fundamentally obscures the biological reality of complex tissues. As David Adams et al. noted in Genome 2011, bulk techniques fail to account for the stochasticity of gene expression and the diversity of clonal evolution, providing a blurred signal that masks individual cell variations.
The Problem of the Average Signal
Imagine trying to understand the dynamics of a crowded city by measuring only the average temperature and noise level of the entire metropolitan area. You might detect a heatwave, but you would miss a fire in a single building or a quiet conversation in a park.
In biological terms, bulk RNA sequencing provides an averaged view of gene expression that hides the specific contributions of individual cell types to disease. Toni Delorey et al. demonstrated this in COVID-19 tissue atlases reveal SARS-CoV-2 pathology and cellular targets, where they found that bulk analysis prevented researchers from identifying which specific cells were infected by SARS-CoV-2. Similarly, Shulan Tian et al. argued in Genomic profiling reveals molecular heterogeneity in patients with Richter transformation (RT) and chronic lymphocytic leukemia (CLL) that this averaged view obscures the precise crosstalk between tumor cells and their immune microenvironment.
The Hidden Minority
The most critical limitation of bulk sequencing is its inability to detect rare but clinically significant cell populations. In cancer, for instance, a small subset of "stem-like" cells may drive resistance to treatment or tumor aggressiveness. Nathan Palmer et al. highlighted in A gene expression profile of stem cell pluripotentiality and differentiation is conserved across diverse solid and hematopoietic cancers that population-level measurements cannot identify these rare, high-plasticity sub-populations.
This "resolution gap" has profound implications for clinical outcomes:
- Treatment Resistance: Xue Lin et al. found that tamoxifen resistance in breast cancer is likely driven by the selection of rare stem-like cells, a process that bulk sequencing cannot distinguish from uniform changes across the whole tumor population in Integrative analyses of gene expression and DNA methylation profiles in breast cancer cell line models of tamoxifen-resistance indicate a potential role of cells with stem-like properties.
- Immune Evasion: In liver cancer, Guohe Song et al. used single-cell resolution to identify specific pro-tumor macrophage subsets that bulk analysis simply could not capture in Global immune characterization of HBV/HCV-related hepatocellular carcinoma identifies macrophage and T-cell subsets associated with disease progression.
- Genetic Counseling: August Yue Huang et al. showed that bulk methods fail to determine the cellular distribution of low-frequency mosaic mutations, which are critical for understanding disease transmission in Postzygotic single-nucleotide mosaicisms in whole-genome sequences of clinically unremarkable individuals.
Obscured Cellular Conversations
Beyond identifying cell types, the resolution gap prevents us from seeing how cells interact. In the brain, for example, a change in gene expression could be a primary defect in a neuron or a secondary, compensatory response from a neighboring astrocyte. Ying Sun et al. noted in RNA-seq co-expression network analysis reveals anxiolytic behavior of mice with Efnb2 knockout in parvalbumin+ neurons that bulk analysis cannot distinguish between these possibilities.
This lack of granularity extends to developmental biology. Weiguo Feng et al. observed that bulk analysis of anatomical structures fails to separate distinct tissue layers, such as the surface ectoderm and the underlying mesenchyme, during facial development in Spatial and Temporal Analysis of Gene Expression during Growth and Fusion of the Mouse Facial Prominences. Similarly, the authors of Molecular signatures of cortical expansion in the human foetal brain argued that transitioning to single-cell transcriptomics is required to identify the precise cellular drivers of regional brain growth.
Bridging the Resolution Gap
To overcome these limitations, the field is shifting toward single-cell technologies and more sophisticated computational models. Yida Wang emphasizes in Analyzing single-cell transcriptome sequencing data reveals immunological mechanisms of microglia-associated diseases that dissecting the roles of specific cells like microglia at a single-cell level is the only way to truly understand their contribution to neurological pathogenesis.
New statistical frameworks are also being developed to extract more meaning from existing data. For example, Wenmin Zhang et al. proposed Gene Score Regression to better interpret genetic associations by moving toward cell-type-specific expression data in Partitioning gene-based variance of complex traits by gene score regression. As these high-resolution tools become standard, the "chaos" of data integration—as described by David Adams et al.—will increasingly reveal the precise cellular mechanisms that bulk sequencing once masked.