RNA-Seq vs scRNA-Seq: Choosing the Right Tool

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Introduction to RNA-Seq and scRNA-Seq

RNA sequencing (RNA-seq) has transformed how we study gene expression, moving from targeted, low-throughput methods like quantitative PCR and microarrays to genome-wide, unbiased quantification of transcripts. The core principle remains the same across all RNA-seq variants: convert RNA into complementary DNA (cDNA), fragment it, ligate adapters, amplify, and sequence. What differs—and what determines the biological questions you can answer—is the input material and the resolution at which you measure expression.

What is Bulk RNA-Seq?

Bulk RNA-seq measures the average gene expression across a population of thousands to millions of cells. You lyse all cells together, extract total RNA, and sequence the resulting cDNA library as a single sample. The output is a single gene expression value per gene per sample, representing the population mean. This approach is ideal for comparing conditions—treated versus untreated, diseased versus healthy, different developmental time points—where the tissue or cell population is relatively homogeneous or where you are explicitly interested in the average response.

Bulk RNA-seq has been the workhorse of transcriptomics since its introduction around 2008. It is mature, well-standardized, and supported by a vast ecosystem of analysis tools. The data are quantitative and reproducible, and the cost per sample is modest compared to single-cell approaches. However, the fundamental limitation is that you lose all information about cellular heterogeneity. If you sequence a tumor biopsy, you get the average expression of cancer cells, stromal cells, immune infiltrates, and endothelial cells mixed together. A gene that is strongly upregulated in 5% of cells may appear unchanged, or even downregulated, in the bulk measurement.

What is Single-Cell RNA-Seq?

Single-cell RNA-seq (scRNA-seq) measures gene expression in individual cells. Each cell is isolated, lysed, and its mRNA is reverse-transcribed with a unique cell barcode, so that after sequencing, every transcript can be traced back to its cell of origin. The output is a large matrix where rows are genes, columns are cells, and each entry is the number of transcripts detected for that gene in that cell.

The resolution gain is transformative. Instead of an average, you obtain a distribution. You can identify distinct cell types, discover rare subpopulations, infer developmental trajectories, and study cell-to-cell variability in gene expression. The trade-off is substantial: scRNA-seq is more expensive per cell, suffers from technical noise (dropout, where a gene expressed in a cell is not detected), and requires considerably more complex bioinformatics for analysis.

The fundamental difference, then, is resolution versus breadth. Bulk RNA-seq gives you a precise, quantitative average at lower cost. scRNA-seq gives you a noisy but high-resolution view of individual cells. Choosing between them is not about which is "better" but about which matches your biological question.

Methodological Workflow Comparison

The experimental workflows for bulk and single-cell RNA-seq diverge at the very first step—sample preparation—and remain distinct through library construction. Understanding these differences is essential for troubleshooting and for interpreting the resulting data.

Sample Preparation and Cell Isolation

Bulk RNA-seq begins with tissue homogenization or cell lysis. For tissue, you typically snap-freeze in liquid nitrogen and grind to a powder, or homogenize directly in a chaotropic lysis buffer such as guanidinium thiocyanate (e.g., TRIzol reagent). Total RNA is then extracted using phenol-chloroform phase separation or silica column-based kits. The RNA is treated with DNase I to remove genomic DNA contamination. Quality is assessed by measuring the RNA Integrity Number (RIN) on a Bioanalyzer; a RIN above 7 is generally acceptable for standard bulk RNA-seq, though lower-quality input can be used with appropriate library prep kits. mRNA is typically enriched either by poly(A) selection using oligo-dT magnetic beads or by ribosomal RNA (rRNA) depletion using probe-based methods (e.g., Ribo-Zero). Poly(A) selection captures messenger RNA and long non-coding RNA with poly(A) tails, while rRNA depletion retains non-polyadenylated transcripts and is preferred for degraded samples or when studying non-coding RNA.

scRNA-seq requires a single-cell suspension. For solid tissues, this means enzymatic dissociation—commonly using collagenase, dispase, or trypsin—followed by mechanical trituration. This step is critical and often the source of batch effects and technical artifacts. Prolonged enzymatic digestion induces stress responses and alters gene expression, particularly for immediate early genes like FOS, JUN, and heat shock proteins. Dead cells must be removed, typically by fluorescence-activated cell sorting (FACS) or magnetic bead-based depletion, because dead cells have high ambient RNA content and produce low-quality libraries. The resulting suspension must be at a precise concentration, usually 700–1,200 cells/µL for droplet-based platforms like 10x Genomics Chromium.

The choice of scRNA-seq platform determines the cell isolation method. Plate-based methods (Smart-seq2, SMART-Seq) sort single cells into individual wells using FACS, allowing visual inspection and full-length transcript coverage. Droplet-based methods (10x Genomics, Drop-seq, inDrop) encapsulate single cells in nanoliter-scale droplets with barcoded beads, enabling massively parallel capture of thousands to tens of thousands of cells per run. The trade-off is cost and sensitivity: plate-based methods are more sensitive and capture full-length transcripts, but are limited to hundreds or low thousands of cells; droplet methods are less sensitive per cell but provide much higher cell throughput.

Library Construction

Bulk RNA-seq library preparation follows a standardized protocol. After mRNA enrichment, the RNA is fragmented, typically by heating to 94°C in the presence of divalent cations (magnesium) for 2–8 minutes to achieve fragments of 200–300 nucleotides. First-strand cDNA synthesis uses reverse transcriptase (e.g., SuperScript II or III) with random hexamers. Second-strand synthesis replaces the RNA template with DNA using RNase H and DNA polymerase I. The double-stranded cDNA is then end-repaired, A-tailed, and ligated to sequencing adapters containing the P5 and P7 flow cell binding sites and an index sequence for multiplexing. The library is amplified by PCR for 10–15 cycles, and size-selected (typically 300–500 bp) using AMPure XP beads. The final library is quantified by qPCR or fluorometry and sequenced.

scRNA-seq library construction is fundamentally different. The key innovation is the addition of a cell barcode and a unique molecular identifier (UMI) during reverse transcription. In the 10x Genomics Chromium platform, a single cell and a gel bead are co-encapsulated in a droplet. The gel bead carries oligonucleotides with a poly(dT) sequence to capture mRNA, a 16-nucleotide cell barcode (shared by all transcripts from that cell), and a 10–12-nucleotide UMI (unique to each mRNA molecule). Reverse transcription occurs inside the droplet, producing cDNA that carries both the barcode and UMI. After breaking the droplets, the cDNA is amplified by PCR, and the library is constructed by enzymatic fragmentation and adapter ligation, similar to bulk RNA-seq. The crucial difference is that the UMI allows you to collapse PCR duplicates: if you see the same UMI and gene multiple times, it represents a single original mRNA molecule, not multiple transcripts. This eliminates PCR amplification bias and enables absolute quantification of transcript counts per cell.

For plate-based methods like Smart-seq2, the entire transcript is reverse-transcribed and amplified, but there is no UMI. Quantification relies on read counts, which are more susceptible to PCR duplication bias. Smart-seq2 provides full-length coverage, enabling isoform detection and allele-specific expression, but at the cost of higher per-cell expense and lower throughput.

Sequencing Platforms

Bulk RNA-seq libraries are sequenced on Illumina platforms (HiSeq, NovaSeq, NextSeq). Paired-end reads of 100–150 bp are standard, providing better mapping to the transcriptome and enabling isoform detection. The required depth depends on the organism and the goal: for differential expression in human or mouse, 20–40 million read pairs per sample is typical; for allele-specific expression or isoform quantification, 60–100 million may be needed. The read length matters: 50 bp single-end reads are sufficient for gene-level quantification but inadequate for isoform-level analysis.

scRNA-seq libraries are also sequenced on Illumina platforms, but the configuration is different. Because the cell barcode and UMI are read in the first few nucleotides of read 1, the read structure is asymmetric. For 10x Genomics, read 1 is 28 bp (containing the barcode and UMI), and read 2 is 90–150 bp (the cDNA insert). Paired-end sequencing is required, but the insert read is the only one that maps to the transcriptome. The sequencing depth per cell is much lower than for bulk: 20,000–50,000 read pairs per cell is typical, which translates to roughly 2,000–5,000 unique transcripts detected per cell. This is a deliberate trade-off: you sacrifice per-cell depth to achieve high cell throughput.

Data Output and Analysis Differences

The data generated by bulk and single-cell RNA-seq are structurally different, and this dictates entirely different bioinformatics pipelines.

Bulk RNA-Seq Data Analysis

Bulk RNA-seq produces a count matrix where rows are genes and columns are samples (typically 3–10 replicates per condition). The analysis pipeline is well-established:

1. Quality control: FastQC for per-base quality scores, adapter trimming with Trimmomatic or cutadapt.
2. Alignment: Spliced aligners like STAR or HISAT2 map reads to the reference genome. Alternatively, pseudo-alignment tools like Salmon or kallisto quantify transcript abundance directly without full alignment.
3. Quantification: FeatureCounts or HTSeq-count generate gene-level count matrices.
4. Normalization: Methods like DESeq2's median-of-ratios or edgeR's trimmed mean of M-values (TMM) account for differences in library size and RNA composition.
5. Differential expression: DESeq2, edgeR, or limma-voom fit a negative binomial or voom-transformed linear model to identify genes with statistically significant expression changes between conditions.
6. Downstream analysis: Gene ontology enrichment, pathway analysis (GSEA, KEGG), and visualization (volcano plots, heatmaps).

The statistical framework is mature because bulk RNA-seq has high replication and low noise. The negative binomial distribution models the count data well, and with 3–5 biological replicates, you can reliably detect 1.5–2-fold expression changes.

scRNA-Seq Data Analysis

scRNA-seq produces a count matrix where rows are genes and columns are cells—typically 5,000–50,000 cells per sample. The analysis pipeline is more complex:

1. Quality control: Filter cells by the number of detected genes (typically 200–5,000), total UMI counts, and the percentage of mitochondrial reads (high mitochondrial content indicates dying cells; >20% is usually discarded).
2. Normalization: Counts are normalized per cell by library size (CPM or log-transformed), but the high dropout rate requires specialized methods like scran's deconvolution or SCTransform, which models UMI counts using a regularized negative binomial regression.
3. Feature selection: Identify highly variable genes (HVGs), typically 2,000–3,000, to reduce dimensionality and focus on informative genes.
4. Dimensionality reduction: Principal component analysis (PCA) on the HVGs, then t-SNE or UMAP for visualization in two dimensions.
5. Clustering: Graph-based clustering (Louvain or Leiden algorithm) partitions cells into groups with similar expression profiles, corresponding to cell types or states.
6. Marker gene identification: Differential expression between clusters identifies marker genes (e.g., CD3D for T cells, CD79A for B cells, LYZ for monocytes) to annotate cell types.
7. Downstream analysis: Trajectory inference (Monocle, Slingshot), cell-cell communication (CellChat, NicheNet), and differential expression between conditions within specific cell types.

Dimensionality Reduction and Clustering

The need for dimensionality reduction and clustering is unique to scRNA-seq. With 20,000 genes and 10,000 cells, the data are sparse (often >90% zeros) and high-dimensional. PCA reduces the data to ~50 principal components that capture the major axes of variation. UMAP or t-SNE then projects these components into two dimensions for visualization, preserving local structure. Clustering on the PCA-reduced space using the Leiden algorithm identifies groups of cells with similar expression profiles. The resolution parameter controls the number of clusters: higher resolution yields more, smaller clusters.

A critical point is that clustering is unsupervised and exploratory. The clusters you obtain depend on the number of HVGs, the number of PCA components, the clustering resolution, and the normalization method. Different parameter choices can yield different cluster numbers, and biological interpretation requires careful validation with marker genes and, ideally, orthogonal methods like flow cytometry or immunohistochemistry.

Biological Questions Addressed by Each Method

The choice between bulk and single-cell RNA-seq should be driven by the biological question, not by technological novelty. Each method answers fundamentally different types of questions.

Bulk RNA-Seq Applications

Bulk RNA-seq is the method of choice when you need precise, quantitative measurements of gene expression changes across conditions. Typical applications include:

• Differential expression between conditions: Comparing wild-type versus knockout, treated versus untreated, or time-course experiments. The statistical power to detect subtle changes (1.5-fold) with modest replication is a major advantage.
• Isoform-level analysis: With paired-end reads and sufficient depth, bulk RNA-seq can quantify alternative splicing events, identify novel isoforms, and measure differential isoform usage. scRNA-seq, especially droplet-based methods, lacks the depth and full-length coverage for reliable isoform quantification.
• Allele-specific expression: Detecting allele-specific expression requires high sequencing depth and the ability to distinguish maternal and paternal transcripts. Bulk RNA-seq with 60–100 million reads per sample is well-suited for this.
• Tissue-level transcriptomic profiling: When the tissue is homogeneous (e.g., liver, where hepatocytes dominate), bulk RNA-seq provides a clean, reproducible measurement without the cost and complexity of single-cell approaches.
• Clinical and diagnostic applications: Bulk RNA-seq is more practical for clinical samples where cell dissociation is not feasible (e.g., formalin-fixed paraffin-embedded tissue) and where cost and turnaround time are critical.

scRNA-Seq Applications

scRNA-seq is the method of choice when heterogeneity is the question. Typical applications include:

• Cell type identification and annotation: Discovering novel cell types or states within a complex tissue. For example, scRNA-seq of the brain has revealed dozens of distinct neuronal and glial subtypes that were invisible to bulk profiling.
• Developmental trajectories: Inferring differentiation pathways from progenitor to mature cells. Trajectory inference algorithms (Monocle, Slingshot) order cells along a pseudotime axis based on transcriptional similarity, reconstructing continuous developmental processes.
• Cell-cell communication: Inferring ligand-receptor interactions between cell types using tools like CellChat or NicheNet. This requires knowing which cell types are present and their expression of signaling molecules.
• Rare cell populations: Identifying and characterizing rare cells, such as circulating tumor cells or tissue-resident immune cells, that constitute <1% of the tissue. Bulk RNA-seq would completely miss these populations.
• Cell cycle and state heterogeneity: Studying cell-to-cell variability in gene expression, such as the continuum of cell cycle states or the heterogeneity of response to a stimulus within a seemingly uniform population.

Technical Challenges and Limitations

Both methods have technical limitations that must be understood to avoid misinterpretation.

Dropout and Noise in scRNA-Seq

The most significant limitation of scRNA-seq is dropout: the failure to detect a transcript that is actually present in the cell. Because each cell contains only ~10–20 pg of total RNA, and the reverse transcription and amplification efficiency is <50%, many transcripts are missed. The result is a sparse count matrix with a high proportion of zeros. Dropout is not random—it is more likely for lowly expressed genes. This creates a technical noise floor that limits the detection of low-abundance transcripts. For a gene expressed at 1 copy per cell, you might detect it in only 10–20% of cells that actually express it. This is why scRNA-seq is not the right tool for detecting lowly expressed genes; bulk RNA-seq, with its pooled input, provides far more sensitive detection.

The UMI design mitigates amplification bias but does not address dropout. The only way to improve detection sensitivity is to increase sequencing depth per cell, but this has diminishing returns—beyond ~50,000 reads per cell, the number of new genes detected plateaus.

Batch Effects in Both Methods

Batch effects—systematic technical variation between samples processed on different days, by different operators, or in different sequencing runs—affect both bulk and single-cell RNA-seq. In bulk RNA-seq, batch effects are managed by experimental design (blocking, randomization) and statistical correction (ComBat, RUVseq). In scRNA-seq, batch effects are more severe because each sample is processed through cell isolation, library preparation, and sequencing, and even minor differences in dissociation time or reagent lot can shift the transcriptome. The standard approach is to use tools like Harmony, Seurat's integration, or Scanorama to align cells across batches. These methods identify shared cell types across batches and correct for technical variation while preserving biological differences.

A critical design principle for scRNA-seq is to ensure that each biological condition is represented in every batch (e.g., use multiplexed samples with lipid-tagged antibodies or hash-tagged oligos to pool samples before processing). This allows batch correction to distinguish technical from biological variation.

Cost and Throughput Considerations

The cost difference is substantial. Bulk RNA-seq library preparation costs $50–150 per sample, and sequencing at 30 million reads costs $100–300. For a typical experiment with 3 replicates per condition and 2 conditions, the total cost is $1,000–2,000. scRNA-seq is considerably more expensive. The 10x Genomics Chromium platform requires a minimum of ~$3,000–4,000 per run for reagents, and each run processes 1–8 samples. Sequencing adds another $1,000–3,000 per run. For a study with 4 samples, expect to spend $5,000–10,000. Plate-based methods like Smart-seq2 are even more expensive per cell ($1–5 per cell) but are used for smaller numbers of cells.

The cost per informative unit is also different. Bulk RNA-seq gives you one measurement per sample; scRNA-seq gives you thousands of measurements per sample. If your question requires identifying rare cell types, the per-cell cost is justified. If your question is about average expression changes, bulk RNA-seq is far more cost-effective.

Choosing Between Bulk and Single-Cell RNA-Seq

The decision framework below summarizes the key considerations.

When to Use Bulk RNA-Seq

Use bulk RNA-seq when:

• Your tissue or cell population is relatively homogeneous, or you are explicitly interested in the average response.
• You need to detect small expression changes (1.5-fold) with statistical confidence.
• You need isoform-level or allele-specific information.
• You have limited RNA (e.g., from laser-capture microdissection) or degraded RNA (FFPE samples).
• Your budget is limited and you need many replicates or many conditions.
• You are working with a model system where you can purify the cell type of interest (e.g., FACS-sorted populations).

When to Use scRNA-Seq

Use scRNA-seq when:

• Your tissue is heterogeneous and you need to identify cell types or states.
• You are studying developmental processes or differentiation trajectories.
• You need to characterize rare cell populations.
• You want to study cell-to-cell variability in response to a stimulus.
• You need to infer cell-cell communication networks.
• You have a well-optimized cell dissociation protocol and can generate a high-quality single-cell suspension.

Hybrid Approaches

A powerful strategy is to combine both methods in a single study. For example, use scRNA-seq to identify cell types and their marker genes, then use bulk RNA-seq (or targeted qPCR) on FACS-sorted populations to validate and quantify expression changes with higher sensitivity. Alternatively, use bulk RNA-seq for a broad screen across many conditions, then use scRNA-seq to dissect the heterogeneity of the most interesting condition. This hybrid approach leverages the strengths of each method while mitigating their weaknesses.

Common Pitfalls and Best Practices

Pitfalls in Experimental Design

Insufficient replication in bulk RNA-seq: Using only 2 replicates per condition is a common mistake. The negative binomial model in DESeq2 requires at least 3 replicates to estimate dispersion reliably. With 2 replicates, you will have high false-positive rates and low power. Use 3–5 biological replicates as a minimum.

Ignoring cell dissociation artifacts in scRNA-seq: Enzymatic dissociation induces transcriptional stress responses. Genes like FOS, JUN, EGR1, and heat shock proteins (HSPA1A, HSP90AB1) are upregulated within minutes of dissociation. This can confound cell type identification and differential expression. Mitigate this by optimizing dissociation time, using cold-active proteases, or adding actinomycin D to block transcription during dissociation.

Not accounting for ambient RNA in scRNA-seq: When cells are lysed during droplet encapsulation, their mRNA is released into the suspension and captured by beads in droplets containing other cells. This creates background contamination. Tools like SoupX or CellBender can estimate and remove ambient RNA, but the best practice is to minimize cell lysis and use a high-viability suspension (>90%).

Pitfalls in Data Analysis

Using bulk RNA-seq normalization for scRNA-seq: Standard bulk normalization methods like DESeq2's median-of-ratios assume that most genes are not differentially expressed between samples. In scRNA-seq, the high dropout rate and the fact that different cell types express different gene sets violate this assumption. Use scRNA-seq-specific methods like SCTransform or scran.

Over-clustering or under-clustering in scRNA-seq: The resolution parameter in Leiden clustering is arbitrary. Too high a resolution creates spurious clusters that split a single cell type; too low merges distinct types. Validate clusters by checking marker gene expression and by using cluster stability metrics.

Ignoring the difference between technical and biological zeros: In scRNA-seq, a zero count can mean either the gene is not expressed or the transcript was missed due to dropout. Do not interpret zeros as biological absence without validation.

Best Practices for Reproducibility

• Document all parameters: Record the exact software versions, alignment parameters, normalization methods, and clustering resolutions. Use a workflow manager like Snakemake or Nextflow.
• Include appropriate controls: For bulk RNA-seq, include a reference RNA sample across batches. For scRNA-seq, include a cell line or PBMC reference to assess batch effects.
• Validate with orthogonal methods: Confirm key findings from scRNA-seq with flow cytometry, immunohistochemistry, or qPCR.
• Use the Sequencing Coverage guidelines: For bulk RNA-seq, 20–40 million reads per sample; for scRNA-seq, 20,000–50,000 reads per cell. Do not over-sequence beyond the point of diminishing returns.
• Follow established Library Prep in Sequencing protocols: Deviations from manufacturer protocols are a common source of batch effects.

Frequently Asked Questions

What is the main difference between RNA-seq and scRNA-seq?

Bulk RNA-seq measures the average gene expression across a population of cells, producing one expression value per gene per sample. scRNA-seq measures gene expression in individual cells, producing a count matrix with one column per cell. The fundamental difference is resolution: bulk gives a population average, while single-cell captures cellular heterogeneity.

When should I use bulk RNA-seq instead of scRNA-seq?

Use bulk RNA-seq when you need precise, quantitative measurements of average expression changes between conditions, when your tissue is homogeneous, when you need isoform-level or allele-specific information, or when your budget and sample availability are limited. Bulk RNA-seq is also preferred for clinical samples where cell dissociation is not feasible.

Why is scRNA-seq more expensive than bulk RNA-seq?

scRNA-seq requires specialized microfluidic or droplet-based equipment, barcoded beads, and higher per-sample reagent costs. The sequencing cost is also higher because you need to sequence barcodes and UMIs in addition to the cDNA insert. For a comparable number of biological samples, scRNA-seq costs 5–20 times more than bulk RNA-seq.

Can scRNA-seq detect lowly expressed genes?

Poorly. The dropout rate in scRNA-seq is high, particularly for lowly expressed genes. A gene expressed at 1 copy per cell may be detected in only 10–20% of cells. Bulk RNA-seq, which pools RNA from thousands of cells, provides far more sensitive detection of low-abundance transcripts. If your gene of interest is lowly expressed, use bulk RNA-seq or validate scRNA-seq findings with qPCR.

What are the common bioinformatics tools for scRNA-seq analysis?

The most widely used tools are Seurat (R package) and Scanpy (Python). Both provide comprehensive pipelines for quality control, normalization, dimensionality reduction, clustering, and marker gene identification. Other tools include Monocle for trajectory inference, Harmony for batch correction, and CellChat for cell-cell communication analysis.

How do I handle batch effects in scRNA-seq data?

The best approach is experimental: design your study so that each biological condition is represented in every batch, ideally by multiplexing samples with hashtag antibodies or lipid-tagged barcodes. For computational correction, use Harmony, Seurat's integration functions, or Scanorama. These methods identify shared cell types across batches and correct for technical variation while preserving biological differences.

Is it possible to do both bulk and single-cell RNA-seq in one study?

Yes, and this is often the best approach. A common strategy is to use scRNA-seq to identify cell types and their marker genes, then use bulk RNA-seq on FACS-sorted populations to validate and quantify expression changes with higher sensitivity. Alternatively, use bulk RNA-seq for a broad screen across many conditions, then use scRNA-seq to dissect the heterogeneity of the most interesting condition.

Key Takeaways

• Bulk RNA-seq measures population-averaged gene expression and is best for differential expression, isoform analysis, and clinical samples; scRNA-seq measures individual cell expression and is best for cell type identification, heterogeneity, and trajectories.
• The experimental workflows diverge at cell isolation: bulk requires RNA extraction from pooled cells, while scRNA-seq requires a viable single-cell suspension and barcoded reverse transcription.
• scRNA-seq data are sparse and noisy due to dropout, requiring specialized normalization, dimensionality reduction, and clustering methods; bulk RNA-seq data are analyzed with mature differential expression pipelines.
• scRNA-seq is 5–20 times more expensive than bulk RNA-seq and requires more complex bioinformatics, but provides resolution that bulk cannot achieve.
• Batch effects affect both methods but are more severe in scRNA-seq; design experiments with balanced batches and use computational correction tools.
• A hybrid approach—using scRNA-seq for discovery and bulk RNA-seq for validation—leverages the strengths of both methods.
• Always validate scRNA-seq findings with orthogonal methods, and document all analysis parameters for reproducibility.

Related Clinical & Scientific Guides

• MAPK Pathway: Mechanism, Function, and Clinical Relevance
• Mammalian Cell Culture Bioreactors: A Practical Guide
• Nucleotide Formation: Biosynthesis and Assembly of DNA/RNA Building Blocks