“Bulk” RNA-seq measures the expression of all genes, on average, across a population of cells. It is the reference tool for comparing transcriptional programmes between conditions, patient groups or treatments — robust, sensitive, economical, applicable to clinical cohorts as well as preclinical models. Where single-cell resolves heterogeneity cell by cell, bulk favours depth, statistical power and cost.
Principle and workflow
Total RNA is extracted, then a sequencing library is prepared — poly-A RNA selection, or ribosomal-RNA depletion (which also captures non-coding RNAs and tolerates degraded RNA or FFPE samples), or 3′ barcoded methods such as BRB-seq, which multiplex many samples at low cost. The fragments are sequenced, aligned to a reference genome, and reads counted per gene give an expression measure. Differential analysis (DESeq2, edgeR) then compares conditions from biological replicates: DESeq2 normalises library sizes and composition, models variability and remains robust for weakly expressed genes. Pathway enrichment (GSEA) and signature reading follow.
Variants and options
The protocol comes in variants according to the objective: poly-A or total RNA (ribosomal depletion) RNA-seq, strand-specific libraries, 3′ barcoded sequencing (BRB-seq, economical multiplexing), small RNA-seq for microRNAs, or pseudo-bulk by aggregating single-cell data. Downstream, deconvolution estimates cell-type fractions from a bulk profile, recovering part of the compositional information.
When and why these techniques
Bulk RNA-seq is the standard when comparing expression programmes between conditions or groups, backed by replicates: the effect of a treatment, a poor-prognosis signature, a pathway repressed in a tumour. It is unbeatable on cohorts (statistical power), defined populations (sorted cells, homogeneous tissue) and archival samples (FFPE), at a cost far below single-cell.
The counterpart of this robustness is a loss of resolution. Being a population average, the measurement masks cellular heterogeneity and rare types, and does not say which type produces a signal. More subtly: an observed variation may reflect either intracellular regulation or a mere change in cell proportions (Simpson’s paradox); resolving the ambiguity requires upstream cell sorting or deconvolution. Power depends on the number of replicates, batch effects must be controlled, and 3′ methods (BRB-seq) forgo isoform information.
Inovarion’s expertise
Inovarion has deployed bulk RNA-seq in oncology and immunology, from patient cohorts to preclinical models. Its publications illustrate this in three contexts: differential expression of CD34+ marrow cells from patients with chronic myelomonocytic leukaemia revealed the repression of retroelements and immune pathways, then their reactivation by an epigenetic combination; transcriptomic analysis of adrenocortical carcinoma linked hypermethylation to immune escape, reversible with a demethylating agent; and BRB-seq profiling of tumour spheroids and patient renal-tumour slices characterised the response to dual kinase inhibition. The choice of library protocol, number of replicates and differential-analysis strategy depends on the tissue interrogated and the resolution required.
See also: Single-cell RNA-seq ; Proteomics & mass spectrometry ; Bioinformatics.
Key publications
- Hidaoui et al. Targeting heterochromatin eliminates chronic myelomonocytic leukemia malignant stem cells through reactivation of retroelements and immune pathways. Communications Biology, 2024. Record → · PubMed
- Kerdivel et al. DNA hypermethylation driven by DNMT1 and DNMT3A favors tumor immune escape contributing to the aggressiveness of adrenocortical carcinoma. Clinical Epigenetics, 2023. PubMed
- Giacosa et al. Cooperative Blockade of CK2 and ATM Kinases Drives Apoptosis in VHL-Deficient Renal Carcinoma Cells through ROS Overproduction. Cancers (Basel), 2021. PubMed