Assessing Microbial Activity from Total RNA-seq in Low-Biomass Tumor Tissue
This repository contains a computational pipeline designed to evaluate whether transcriptionally active microbial signals can be reliably detected in low-biomass lung tumor tissues using total RNA-seq data. Using two independent early-stage NSCLC cohorts, this workflow applies a contamination-aware metatranscriptomic processing strategy, including:
- Host read removal
- Quality control
- Taxonomic classification
- Downstream ecological and statistical analyses
Because tumor RNA-seq datasets are overwhelmingly dominated by human transcripts, distinguishing true microbial activity from technical noise is a major challenge. This pipeline emphasizes:
- Stringent bioinformatic filtering
- Contamination-aware microbial classification
- Reproducibility assessment across cohorts
- Ecological diversity metrics and differential abundance testing
The goal of this work is to:
- Characterize potential microbial profiles in low-biomass tumor tissues.
- Highlight the methodological pitfalls of identifying low-abundance microbial signals in host-focused sequencing experiments.
- Provide a transparent and reproducible framework for researchers studying tumor microbiomes.
- Reproducible pipeline for metatranscriptomic analysis
- Handles low-biomass, host-dominated RNA-seq datasets
- Integrates ecological and statistical analyses to assess microbial activity
- Designed for transparency and methodological rigor