Inferring Immune Infiltration from Bulk Transcriptomic Data

You are reading version 1 of this direction, frozen 2026-10-04. The latest version is version 2.

Version 1 of 2502 papers571 conceptsfrozen 2026-10-04

In cancer research and regenerative medicine, a tissue sample is often treated like a "smoothie." When researchers perform bulk RNA sequencing, they grind up a piece of a tumor or an injured organ, resulting in a single measurement of gene expression that averages across millions of different cells. This poses a problem: a high expression of a specific gene might mean the cancer cells are hyperactive, or it might mean the tissue is swarming with immune cells.

To solve this without the high cost of single-cell sequencing or the physical labor of cell sorting, researchers have turned to computational deconvolution. This approach treats the bulk transcriptomic profile as a mathematical mixture and attempts to "unmix" it to reveal the cellular composition of the tumor microenvironment (TME).

The Logic of Digital Deconvolution

The intuition behind this method relies on "reference signatures." Every cell type—whether a CD8+ T cell, a macrophage, or a fibroblast—has a unique transcriptional fingerprint. If we know these fingerprints, we can use algorithms like CIBERSORT or ssGSEA (single-sample gene set enrichment analysis) to estimate the relative abundance of each cell type in a mixed sample.

As Pietro E. Cippà, Andrew P. McMahon, and colleagues demonstrated in A late B lymphocyte action in dysfunctional tissue repair following kidney injury and transplantation, this allows for the tracking of immune cell infiltration patterns across different stages of disease progression. By applying deconvolution to bulk RNA-seq data from kidney biopsies, they could identify a specific "broken" repair loop driven by B lymphocytes without needing to physically isolate the cells at every time point.

From Cell Counts to Prognostic Scores

Once the cellular composition is inferred, researchers can calculate "immune scores" and "stromal scores." These numerical values represent the "hotness" or "coldness" of a tumor—essentially, how well the immune system has successfully infiltrated the cancer's defenses.

In Prognostic value of immune scores in the microenvironment of colorectal cancer, Wenliang Yuan and Sihua Peng showed that these scores are powerful prognostic markers. Patients with high immune scores in colorectal cancer often have "hot" microenvironments with active T-cell recruitment, leading to significantly better survival outcomes. Similarly, Chaoqi Zhang and Nan Sun developed a Costimulatory Molecule-based Signature for lung adenocarcinoma that uses these inferred immune states to predict which patients will respond best to immunotherapy.

Identifying the Molecular Drivers of Infiltration

The most recent frontier in this field is moving from observing the microenvironment to identifying the drivers that shape it. By correlating the expression of specific genes with inferred immune cell densities, researchers can pinpoint which molecules act as "beacons" or "shields" for the tumor.

For instance, Fei Gu and Xin Xing found that the gene TIMELESS actively recruits macrophages to ovarian cancer sites, creating a pro-tumor environment. Other researchers are looking at the "epitranscriptome"—chemical modifications on RNA itself. Chuanxi Yang and colleagues explored how the ac4C regulator NAT10 influences immune infiltration across multiple cancer types, suggesting that the way RNA is modified can dictate how the immune system perceives a tumor.

Beyond Oncology

While primarily used in cancer, this "in silico" profiling is expanding into other areas of acute medicine. In Signatures of inflammation and impending multiple organ dysfunction in the hyperacute phase of trauma, Claudia Cabrera and her team used deconvolution to identify a genomic signature in the blood of trauma patients within just two hours of injury. This early window allows clinicians to predict the trajectory toward organ failure long before systemic inflammation becomes visible through traditional means.

By treating bulk transcriptomics as a rich, deconvolvable signal rather than a simple average, this research direction has turned standard clinical data into a high-resolution map of the cellular interactions driving disease and recovery.

Papers in this version

Showing 60 of 502 papers.

History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 256 papers
  2. Version 1 · 2026-10-04 · 502 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_transcriptomic_immune_deconvolution_ed14f484,
  title = {transcriptomic immune deconvolution},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/transcriptomic-immune-deconvolution/rdr_ed14f4849f4b432e872a8ff5f309a334},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_inferring_immune_infiltration_from_bulk__ed14f484_v1,
  title = {Inferring Immune Infiltration from Bulk Transcriptomic Data},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/inferring-immune-infiltration-from-bulk-transcriptomic-data/rdr_ed14f4849f4b432e872a8ff5f309a334/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}