Macrophage State Transitions in Lung Cancer

Integrating RNA velocity and long-read single-cell transcriptomics to investigate macrophage plasticity in non-small cell lung cancer.

Project Overview

Project Title: Integrating RNA Velocity and Isoform-Level Analysis to Define Macrophage State Transitions in Non-Small Cell Lung Cancer

Research Areas: Computational Biology · Bioinformatics · Single-Cell Transcriptomics · RNA Velocity · Long-Read Sequencing · Cancer Immunology

Tumor-associated macrophages (TAMs) are highly plastic immune cells that contribute to inflammation, antigen presentation, tissue remodeling, and immune suppression in non-small cell lung cancer (NSCLC). Although single-cell RNA sequencing has revealed substantial macrophage heterogeneity, most analyses provide a largely static view of cell populations.

Our goal is to develop an integrated computational framework that connects macrophage state transitions with transcript isoform variation by combining RNA velocity, trajectory inference, and long-read single-cell RNA sequencing. This approach is designed to identify dynamic transcriptional programs associated with tumor-related macrophage states and prioritize candidate transcript isoforms for future mechanistic investigation.

Research Framework

The project integrates two complementary analytical branches:

  • Short-read single-cell RNA sequencing: characterize macrophage heterogeneity, compare tumor-related and non-tumor cell populations, infer potential state transitions, and identify transition-associated genes using RNA velocity and complementary trajectory methods.
  • Long-read single-cell RNA sequencing: quantify known and candidate novel transcript isoforms, investigate isoform usage changes, and establish a reproducible SCOTCH-based analysis workflow.

These results are integrated to prioritize transcript isoforms associated with tumor-enriched macrophage programs and inferred cellular transitions.

Figure 1. Computational framework integrating short-read single-cell analysis of macrophage states and transitions with long-read transcript isoform analysis.

Research Objectives

  1. Reconstruct macrophage state transitions in NSCLC using single-cell clustering, RNA velocity, pseudotime, and tumor-versus-peritumoral comparisons.
  2. Characterize transcript isoform variation using long-read single-cell RNA sequencing and SCOTCH, including isoform quantification and candidate isoform switching.
  3. Integrate cellular dynamics with transcript-level changes to prioritize candidate isoforms associated with tumor-enriched macrophage states and transition-related programs.

Because matched short-read and long-read NSCLC macrophage datasets remain limited, cross-dataset integration is interpreted as hypothesis-generating rather than definitive evidence of isoform regulation during a specific cellular transition.

Key Preliminary Results

Using the publicly available NSCLC single-cell RNA-seq dataset GSE131907, we analyzed 34,420 monocyte/macrophage-lineage cells, including monocytes, monocyte-derived macrophages, alveolar macrophages, and pleural macrophages.

Our preliminary analysis shows:

  • Substantial macrophage heterogeneity: dimensionality reduction and high-resolution Leiden clustering resolved multiple transcriptionally distinct macrophage states.
  • Tumor-associated shifts in macrophage composition: tumor-related samples were enriched for monocyte-derived macrophages, whereas non-tumor samples were dominated by alveolar macrophages.
  • Distinct functional programs: marker-gene analysis identified inflammatory (IL1B, CXCL8, TNF), antigen-presenting (HLA-DRA), lipid-associated (APOE, SPP1, TREM2), and extracellular-matrix-remodeling (MMP9, VCAN, TIMP1) signatures.

These findings provide a foundation for identifying potential macrophage state transitions and linking dynamic cell-state programs with transcript-level regulation.

Figure 2. Single-cell characterization of the NSCLC macrophage compartment, illustrating macrophage heterogeneity, published subtype annotations, higher-resolution Leiden clusters, and tumor-related versus non-tumor distributions.
Figure 3. Tumor-associated differences in macrophage composition and higher-resolution cluster distributions across tissue contexts.

Computational Methods

Methods and technologies: Scanpy, scVelo, RNA velocity, pseudotime and trajectory analysis, Leiden clustering, marker-gene analysis, long-read single-cell RNA sequencing, SCOTCH, transcript isoform analysis, Python, version-controlled workflows, and high-performance computing.

Research Impact

This project aims to move beyond static descriptions of macrophage heterogeneity by linking inferred cellular dynamics with transcript isoform regulation. The expected outcomes include a macrophage-focused transition map, a reproducible long-read single-cell isoform analysis workflow, and a prioritized set of candidate transcript isoforms associated with tumor-related macrophage programs.

More broadly, the computational framework can support future mechanistic studies and be adapted to investigate dynamic cellular processes in other disease systems.

Project Information

Principal Investigator: Xiuquan Wang, Ph.D.
Institution: Tougaloo College
Research Program: Computational Biology and Bioinformatics
Funding: MS-INBRE Project Development Grant
Project Period: 2026–2028

Figures and findings presented here are from ongoing research. Additional computational analyses and experimental validation are planned.

References