Written and maintained by CASRAI Editorial Board
Published
On this page
The standard RNA-seq steps encompass a sequential ten-stage workflow spanning wet-bench preparation and bioinformatic analysis: total RNA isolation, fluorometric quantification, capillary electrophoresis quality assessment (RIN or DV200), mRNA enrichment or ribosomal RNA depletion, enzymatic fragmentation and stranded cDNA synthesis, adapter ligation, PCR indexing, molar library pooling, high-throughput sequencing-by-synthesis, and raw read processing through alignment and differential expression.
Each phase requires stringent quality control checkpoints to avoid compounding technical artifacts across downstream data processing. Understanding the biochemical rationale behind every step—such as deoxyuridine triphosphate (dUTP) second-strand incorporation for stranded transcription orientation or empirical titration of magnetic solid-phase reversible immobilization (SPRI) beads—ensures that sequencing budgets generate reproducible, publication-grade transcriptome measurements.
What you need: reagents, equipment, and consumables
Executing an RNA-seq workflow from biological tissue through sequencing-ready libraries requires dedicated molecular biology instrumentation, nuclease-free reagents, and high-performance computational infrastructure. Because environmental ribonuclease (RNase) enzymes rapidly degrade RNA molecules, all pre-PCR benchwork must take place in an RNase-free environment with designated pipettes, filter tips, and decontamination sprays.
- RNA isolation and cleanup reagents: Guanidinium thiocyanate-phenol-chloroform reagent (such as TRIzol) or silica-membrane spin column extraction kits (such as the Qiagen RNeasy Mini Kit), RNase-free DNase I enzyme sets for on-column or in-solution genomic DNA digestion, molecular biology-grade absolute ethanol (prepared fresh as 80% and 70% v/v working solutions in nuclease-free water), and nuclease-free water certified free of detectable RNase and DNase activity.
- Quality control consumables and reagents: Qubit RNA High Sensitivity (HS) and Broad Range (BR) Assay Kits with thin-wall 0.5 mL optical tubes for fluorometric quantification, Agilent RNA 6000 Nano and Pico Kits (chips, gel-dye matrix, and marker) for Bioanalyzer analysis, or RNA ScreenTape and Reagents for 4200 TapeStation capillary electrophoresis.
- Library preparation kits: Commercial stranded library preparation chemistry, such as the Illumina Stranded mRNA Prep (Ligation) Kit, the Illumina Stranded Total RNA Prep with Ribo-Zero Plus, or the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina, along with compatible Unique Dual Index (UDI) primer plates (such as Illumina DNA/RNA UD Indexes or NEBNext Multiplex Oligos).
- Purification beads and magnetic hardware: Solid Phase Reversible Immobilization (SPRI) paramagnetic carboxylated beads (such as Beckman Coulter AMPure XP or NEBNext Sample Purification Beads), paired with a 96-well magnetic separation rack or 0.2 mL magnetic tube stand.
- Bench instruments: Microcentrifuge refrigerated to 4°C, calibrated thermal cycler with an adjustable heated lid capable of maintaining 100°C (with a 4°C hold function), vortex mixer, Qubit 4 Fluorometer, and Agilent 2100 Bioanalyzer or TapeStation system.
- Sequencing platform: Next-generation sequencing system (such as an Illumina NovaSeq 6000, NovaSeq X Plus, NextSeq 2000, or MiSeq instrument) equipped with corresponding flow cell cartridges, clustering reagents, and Illumina PhiX Control v3 library spiked at 1% to balance nucleotide diversity.
- Computational environment: 64-bit Linux or Unix workstation or high-performance computing (HPC) cluster node with at least 32 to 64 GB of RAM, multi-core processing architecture, and pre-installed bioinformatics packages including FastQC, fastp or Trimmomatic, STAR aligner, Subread (featureCounts), and an R environment configured with Bioconductor packages DESeq2 and edgeR.
RNA-seq protocol steps: complete end-to-end workflow
The RNA-seq protocol translates cellular transcriptomes into digital read counts through ten precise, interdependent stages. Deviations in temperature, incubation duration, or bead-to-sample ratios at any bench stage directly alter insert size distributions, introduce 3-prime or 5-prime transcript coverage biases, or compromise library yields.
Step 1: Total RNA isolation and sample homogenization
The protocol begins with rapid cellular lysis to stabilize fragile RNA species before endogenous ribonucleases initiate degradation. Biological specimens—such as mammalian cell pellets, snap-frozen tissue biopsies, or plant leaves—are lysed in a guanidinium isothiocyanate-based denaturing solution or specialized lysis buffer. Solid tissues require mechanical disruption using a bead mill homogenizer (e.g., 30 Hz for 2 to 3 minutes with stainless steel beads) or rotor-stator homogenizers on wet ice.
Following homogenization, lysate phase separation or column binding isolates total RNA from cellular proteins, lipids, and polysaccharides. A critical requirement during extraction is enzymatic digestion of co-purified genomic DNA (gDNA). Incubate the sample with RNase-free DNase I for 15 minutes at room temperature (20°C to 25°C), either directly on the silica spin column membrane or in liquid suspension prior to final ethanol precipitation. Elute the purified total RNA in 30 to 50 µL of nuclease-free water, aliquot into low-binding microcentrifuge tubes, and immediately place on ice or store at -80°C.
Step 2: RNA quantification and integrity assessment (RIN and DV200)
Accurate quantification and integrity verification are essential before committing samples to library construction. First, measure purity using a UV-Vis spectrophotometer: pure RNA yields an optical absorbance ratio of A260/A280 ≈ 2.0 and an A260/A230 ≥ 2.0. Absorbance readings below 1.8 indicate contamination by residual phenol, guanidine salts, or proteins. Because spectrophotometry cannot distinguish intact RNA from free nucleotides or degraded fragments, determine actual concentration using a fluorometer (such as the Qubit Fluorometer with the RNA HS Assay), which employs an RNA-selective fluorophore.
Next, evaluate RNA structural integrity by microfluidic capillary electrophoresis on an Agilent 2100 Bioanalyzer using the RNA 6000 Nano or Pico chip. The instrument software calculates the measuring and interpreting RNA integrity number (RIN), an empirical metric on a 1-to-10 scale derived from the 28S and 18S ribosomal RNA peak areas and electropherogram baseline degradation signals. For standard poly(A) mRNA enrichment protocols, samples should exhibit a RIN ≥ 8.0 (with 7.0 representing the absolute lower threshold). For formalin-fixed paraffin-embedded (FFPE) or archival clinical specimens where chemical cross-linking causes severe degradation, RIN values lose discriminatory power; in these cases, evaluate the DV200 metric—the percentage of RNA fragments exceeding 200 nucleotides in length—where a DV200 ≥ 30% indicates sufficient fragment length for total RNA library preparation.
Step 3: mRNA selection or ribosomal RNA depletion
In total cellular RNA, ribosomal RNA (rRNA) species (specifically 28S, 18S, 5.8S, and 5S in eukaryotes) account for 80% to 90% of all molecules, while transfer RNAs (tRNAs) represent another 10% to 15%. Protein-coding messenger RNA (mRNA) constitutes merely 1% to 5% of the total pool. Without targeted enrichment or depletion, sequencing reads will be overwhelmingly consumed by uninformative rRNA sequences.
Researchers must select one of two primary strategies based on input quality and experimental objectives:
- Oligo(dT) selection (poly(A) capture): Paramagnetic polystyrene beads coated with oligo-deoxythymidine (dT) capture the 3-prime polyadenylated tails of mature eukaryotic mRNAs. Total RNA (typically 25 ng to 1,000 ng) is denatured at 65°C for 5 minutes, bound to beads at room temperature, washed to remove unbound non-polyadenylated RNA, and eluted. This process is repeated for two consecutive binding rounds to achieve greater than 98% rRNA clearance. Oligo(dT) selection is highly cost-effective and produces deep coverage across coding exons, but it requires intact RNA (RIN ≥ 8.0) and fails to capture non-polyadenylated regulatory transcripts or prokaryotic RNA.
- Ribosomal RNA depletion: Sequence-specific DNA or RNA probes complementary to cytoplasmic and mitochondrial rRNAs hybridize to target transcripts. The resulting RNA-DNA hybrids are either degraded by catalytic RNase H cleavage or captured and removed via streptavidin magnetic beads. Because rRNA depletion targets specific sequences rather than the 3-prime terminal tail, it preserves partially degraded transcripts, pre-mRNAs, and long non-coding RNAs (lncRNAs), making it the mandatory method for FFPE clinical material, prokaryotic systems, and non-coding RNA investigations.
Step 4: RNA fragmentation and first-strand cDNA synthesis
High-throughput sequencing-by-synthesis instruments require library inserts sized between 200 and 400 base pairs to ensure efficient bridge amplification or exclusion clustering across the flow cell. Purified mRNA or rRNA-depleted RNA undergoes controlled thermal fragmentation in a divalent cation buffer (containing magnesium or zinc ions) at 94°C. In the Illumina Stranded mRNA protocol, fragmentation is performed at 94°C for exactly 8 minutes in Elute, Prime, Fragment 3HC Mix (EPH3). In the NEBNext Ultra II Directional RNA protocol, intact RNA is incubated at 94°C for 15 minutes to target ~200 nt inserts, whereas partially degraded RNA (RIN 2 to 6) is incubated for only 7 to 8 minutes, and severely degraded FFPE RNA requires no thermal fragmentation.
Immediately following fragmentation, random hexamer primers anneal to the RNA fragments as the reaction cools to 4°C. Reverse transcriptase enzyme (such as engineered M-MuLV reverse transcriptase) synthesizes the first-strand complementary DNA (cDNA). Actinomycin D is present in the first-strand reaction mix to inhibit DNA-dependent DNA synthesis while permitting RNA-dependent RNA transcription, preventing spurious second-strand artifacts during the initial polymerization phase.
Step 5: Second-strand synthesis with dUTP for strand specificity
Preserving transcriptional strand orientation is critical for distinguishing antisense regulatory transcripts, identifying novel overlapping genes, and accurately quantifying expression from closely spaced loci on opposing genomic strands. Modern protocols achieve strand specificity through the dUTP second-strand synthesis method.
During second-strand synthesis, DNA Polymerase I and RNase H nick the original RNA strand and synthesize the replacement cDNA strand. Crucially, the deoxynucleotide triphosphate mix replaces deoxythymidine triphosphate (dTTP) with deoxyuridine triphosphate (dUTP). Consequently, the newly synthesized second cDNA strand incorporates dUTP residues throughout its backbone. Following synthesis, double-stranded cDNA is purified using AMPure XP paramagnetic beads (typically at a 1.8x bead-to-sample volume ratio) to remove enzymes, unincorporated dNTPs, and short oligonucleotides.
Step 6: End repair, A-tailing, and adapter ligation
The resulting double-stranded cDNA fragments possess ragged ends with variable 5-prime and 3-prime single-stranded overhangs. The end repair enzyme cocktail (comprising T4 DNA polymerase, Klenow fragment, and T4 polynucleotide kinase) blunts the cDNA ends by filling in 5-prime overhangs and chewing back 3-prime overhangs through 3-prime-to-5-prime exonuclease activity, while phosphorylating the 5-prime termini.
Next, Klenow exo-minus enzyme adds a single non-templated deoxyadenosine monophosphate (A) to the 3-prime ends of the blunt, phosphorylated duplexes (A-tailing). This single-base 3-prime A overhang prevents cDNA fragments from self-ligating and provides a specific base-pairing target for sequencing adapters possessing a single complementary 3-prime deoxythymidine (T) overhang.
Sequencing adapters containing flow cell attachment sequences (P5 and P7) and sequencing primer binding sites (Read 1 and Read 2) are then ligated onto the A-tailed cDNA fragments using T4 DNA ligase at 20°C for 15 minutes. After ligation, double-sided or single-sided SPRI bead cleanups (0.8x to 1.0x volume ratio) remove excess unligated adapters and adapter-dimer complexes, which would otherwise compete for flow cell cluster generation.
Step 7: PCR amplification and dual-index barcode incorporation
Adapter-ligated cDNA libraries are amplified by high-fidelity PCR to achieve sufficient molar mass for sequencing and to incorporate sample-specific indexing barcodes. The PCR master mix incorporates a specialized proofreading DNA polymerase that is chemically engineered to stall when encountering uracil bases. During thermal denaturation and elongation, the dUTP-containing second cDNA strand cannot serve as a template, ensuring that only the original first-strand cDNA (representing the antisense transcript orientation) is replicated.
PCR primers introduce Unique Dual Indexes (UDIs), which append distinct 8-base-pair or 10-base-pair barcode sequences at both the i5 and i7 adapter positions. Using UDIs prevents sample misassignment caused by index hopping—a phenomenon common to patterned flow cells where free index primers migrate between library molecules during pooled exclusion amplification. PCR cycle numbers must be strictly calibrated against starting total RNA input: 10 to 12 cycles are sufficient for 100 ng of input, whereas low-input libraries (10 to 25 ng) require 13 to 15 cycles. Over-amplification introduces PCR duplication bias and non-linear skewing of abundant transcripts. A final 0.8x to 0.9x AMPure XP bead purification removes remaining primers and dNTPs, eluting the completed library in 20 to 30 µL of 10 mM Tris-HCl buffer (pH 8.0 to 8.5).
Step 8: Library quantification, fragment sizing, and molar pooling
Before loading onto a flow cell, completed libraries must be validated for size distribution and precise molar concentration. Analyze 1 µL of the library on an Agilent 2100 Bioanalyzer using a High Sensitivity DNA chip or on a TapeStation using a D1000 ScreenTape. High-quality libraries present as a smooth, unimodal distribution centered between 260 bp and 350 bp. A distinct, sharp peak at 120 bp to 140 bp indicates residual adapter-dimer contamination, which must be eliminated via an additional 0.8x bead cleanup before sequencing.
Quantify library concentrations using a Qubit dsDNA HS Assay for fluorometric measurement, or ideally via quantitative real-time PCR (qPCR) using primers targeting the Illumina P5 and P7 adapter sequences (e.g., KAPA Library Quantification Kit). qPCR measures only those library molecules with functional, full-length adapters capable of forming flow cell clusters. Convert mass concentration (ng/µL) to molarity (nM) using the standard molecular formula:
Molarity (nM) = [Concentration (ng/µL) / (660 g/mol × Average Fragment Length in bp)] × 106
Normalize all multiplexed libraries to a uniform working concentration (typically 2 nM to 4 nM) and pool equal volumes of each normalized library into a single multiplexed sequencing pool.
Step 9: High-throughput sequencing-by-synthesis
Denature the pooled library with freshly prepared dilute sodium hydroxide (0.1 N or 0.2 N NaOH) at room temperature for 5 minutes to generate single-stranded DNA templates, followed by dilution in chilled hybridization buffer (such as HT1 buffer). Spike in a balanced control library (Illumina PhiX Control v3) at a 1% concentration to monitor error rates and provide nucleotide diversity during the initial imaging cycles.
Load the denatured pool into the sequencing cartridge and initiate the run. In modern sequencing platforms (such as the NovaSeq 6000 or NovaSeq X series), single-stranded molecules bind to surface-grafted oligonucleotides on patterned nanowells, followed by exclusion amplification (ExAmp) to generate monoclonal DNA clusters. Sequencing-by-synthesis (SBS) proceeds through cyclical additions of fluorescently labeled reversible terminator dNTPs, with optical imaging recording base incorporation at each cycle. For mammalian transcriptome profiling, standard guidelines established by the ENCODE consortium quality standards recommend paired-end sequencing (such as 2 × 100 bp or 2 × 150 bp) with a minimum depth of ≥ 30 million mapped read pairs per biological replicate to reliably capture alternate splicing events and quantify low-abundance transcripts.
Step 10: Quality control, read alignment, and differential expression
Following sequencing completion, instrument base-calling software generates raw BCL files, which are demultiplexed into compressed FASTQ files per sample using bcl2fastq2 or bcl-convert. The downstream computational pipeline proceeds through four sequential phases:
- Raw read quality assessment: Run FastQC to evaluate per-base Phred quality scores (Q30 ≥ 85%), sequence duplication levels, overrepresented adapter sequences, and GC content distribution. Inspect per-tile sequence quality metrics to detect optical or fluidic anomalies on the flow cell.
- Adapter trimming and filtering: Use fastp, Trimmomatic, or Cutadapt to trim residual adapter sequences, remove low-quality base calls from read termini (Q < 20), and filter out reads shorter than 35 base pairs.
- Splice-aware read alignment: Align trimmed reads to an annotated reference genome (e.g., GRCh38 for human or GRCm39 for mouse) using a splice-aware aligner such as the STAR aligner for splice-aware mapping or HISAT2. Splice-aware alignment is mandatory because eukaryotic mRNA reads span intron-exon junctions. STAR outputs coordinate-sorted BAM files alongside junction mapping tables and alignment summary logs.
- Feature quantification and differential testing: Quantify read overlap against annotated genomic features (GENCODE or Ensembl GTF files) using featureCounts or pseudoaligners like Salmon and kallisto. Import raw count matrices into statistical testing frameworks. When selecting analytical software, review the trade-offs between DESeq2 vs edgeR differential expression tools; both apply empirical Bayes shrinkage to estimate dispersion across biological replicates and model count data using negative binomial generalized linear models. Align experimental comparisons with broader experimental design considerations, ensuring at least three biological replicates per condition to distinguish genuine transcriptional modulation from biological variance.
Sourced decision guide: poly(A) enrichment vs rRNA depletion vs total RNA-seq
Selecting an appropriate library preparation chemistry represents one of the most consequential decisions in experimental planning. Choosing the wrong strategy can lead to severe 3-prime transcript coverage bias, loss of unpolyadenylated regulatory RNAs, or catastrophic waste of sequencing capacity on ribosomal RNA fragments. The dataset table below provides comparative technical specifications across major commercial RNA-seq library preparation kits, detailing input tolerances, fragmentation conditions, and recommended read depths.
The comparative matrix below evaluates the technical operational parameters across the three primary RNA-seq enrichment strategies based on manufacturer specifications and empirical consortium benchmarks.
| Operational Parameter | Oligo(dT) Poly(A) Selection | Ribosomal RNA Depletion | Whole Transcriptome / 3-Prime Tag |
|---|---|---|---|
| Input RNA mass range | 25 ng to 1,000 ng total RNA | 10 ng to 1,000 ng total RNA | 0.5 ng to 500 ng total RNA |
| Minimum RNA integrity (RIN) | RIN ≥ 7.0 (optimal RIN ≥ 8.0) | RIN ≥ 2.0 (compatible with FFPE DV200 ≥ 30%) | RIN ≥ 1.0 to 10.0 (unaffected by degradation) |
| Transcriptome coverage | Exclusively polyadenylated mRNAs and select lncRNAs | Coding mRNAs, pre-mRNAs, lncRNAs, snoRNAs, circRNAs | Only 3-prime terminal mRNA ends (one read per transcript) |
| Intron and pre-mRNA capture | Minimal (under 5% intronic reads in intact cells) | Substantial (15% to 35% intronic and nascent pre-mRNA reads) | Negligible (under 2% intronic reads) |
| Prokaryotic suitability | Unusable (bacterial mRNAs lack stable poly(A) tails) | Mandatory (requires pan-bacterial rRNA depletion probes) | Unusable (requires polyadenylated 3-prime terminus) |
| Recommended read depth | 20 to 30 million mapped reads per sample | 40 to 60 million mapped reads per sample | 5 to 10 million mapped reads per sample |
| Alternative splicing detection | High across all coding exons | High across coding and non-coding exons | None (captures only 3-prime untranslated regions) |
| Primary operational advantage | Maximum sequencing efficiency per dollar for coding genes | Tolerates degraded RNA and captures non-coding transcripts | Lowest sequencing cost and simplified counting analysis |
When working with high-quality eukaryotic cell cultures or flash-frozen animal tissues exhibiting RIN ≥ 8.0, oligo(dT) poly(A) enrichment delivers superior statistical power for differential gene expression. Because sequencing reads are focused strictly on mature mRNA exons, lower total sequencing depth (25 to 30 million reads) is required per library.
Conversely, for clinical biopsy tissues, archival formalin-fixed paraffin-embedded (FFPE) blocks, or research programs investigating circular RNAs (circRNAs) or long intergenic non-coding RNAs (lincRNAs), ribosomal RNA depletion is indispensable. Poly(A) selection on degraded RNA leads to severe 3-prime coverage bias, where only the 3-prime terminal fragments containing the poly(A) tail are captured, leaving 5-prime and middle exons unsequenced. Ribosomal RNA depletion requires higher overall sequencing depth (40 to 60 million reads) because a substantial proportion of reads map to intronic regions of nascent pre-mRNAs.
What can go wrong: troubleshooting common RNA-seq failures
RNA sequencing workflows feature multiple biochemical and fluidic failure modes that manifest during bench handling or post-sequencing bioinformatic quality audits. Identifying root causes early prevents repeating expensive sequencer runs.
- Genomic DNA contamination in library reads: If DNase I treatment is incomplete during extraction, co-purified genomic DNA duplexes will fragment alongside RNA during thermal incubation. In downstream alignment metrics, gDNA contamination manifests as abnormally high intergenic mapping rates (exceeding 15%) and uniform read coverage across introns and unexpressed genomic regions. To resolve this, always implement on-column DNase I digestion followed by an in-solution DNase cleanup step for DNA-rich tissues such as spleen or thymus.
- Severe 3-prime gene body coverage bias: In poly(A)-selected libraries, an elevated ratio of read coverage at transcript 3-prime ends relative to 5-prime ends indicates that the starting RNA was degraded before bead capture. In bioinformatic QC (such as Picard CollectRnaSeqMetrics), an intact sample exhibits a flat, uniform gene body coverage profile between 0.1 and 0.9 normalized transcript positions. If gene body coverage slopes sharply upward toward the 3-prime end, the input RIN was inadequate for poly(A) selection; re-prep the samples using an enzymatic rRNA depletion kit.
- High residual ribosomal RNA reads (exceeding 10%): If post-alignment metrics show excessive read mapping to ribosomal RNA loci (5S, 5.8S, 18S, 28S, 12S, or 16S), the enrichment or depletion phase failed. Primary causes include using expired oligo(dT) or magnetic probe reagents, insufficient bead wash agitation, or carrying residual ethanol into the final elution. In rRNA depletion protocols, probe-to-RNA input ratios must be precisely matched to species-specific rRNA concentrations. Ensure thorough bead pelleting on the magnet before aspirating supernatant washes.
- Adapter-dimer amplification peaks (120 to 140 bp): Adapter dimers form when P5 and P7 adapter oligonucleotides ligate directly to each other without an intervening cDNA insert. Because these short fragments are ~120 to 140 bp in length, they amplify with high kinetic efficiency during PCR and preferentially cluster on sequencing flow cells. An adapter-dimer peak exceeding 1% of total molarity on a Bioanalyzer trace can consume 20% to 50% of total flow cell reads, yielding unusable data. Remediate contaminated libraries by performing an additional 0.8x to 0.9x SPRI bead purification before pooling. For low-input samples (under 10 ng), dilute the adapter stock 1:5 or 1:10 according to manufacturer recommendations.
- Flow cell overclustering and severe Q30 score drops: In Illumina sequencing, loading library pools at excessive molar concentrations causes spatial overlapping of optical clusters. Overclustering prevents the instrument cameras from resolving discrete signal boundaries, leading to focus failures, high error rates, and dramatic drops in Phred quality scores (especially during Read 2). On patterned flow cells (NovaSeq), overclustering manifests as low percentage passing filter (%PF) and index assignment failures. Conversely, underclustering yields low overall read output. Always calibrate loading molarity using qPCR rather than fluorometry alone.
- Technical batch effects confounding biological signal: When sample batches are processed on different calendar days, by different technicians, or using different reagent kit lots, technical batch variation can completely obscure biological differential expression. In exploratory Principal Component Analysis (PCA), batch effects manifest as primary clustering by extraction date rather than experimental treatment. Prevent batch confounding at the experimental design stage by blocking and randomizing samples across extraction batches, library plates, and sequencer flow cell lanes.
When to get professional help
While standard bulk RNA-seq workflows on intact eukaryotic cell cultures can be routinely executed in molecular biology laboratories, specialized experimental conditions demand consultation with institutional genomics core facilities or professional bioinformaticians:
- Degraded clinical and FFPE tissue processing: Handling archival clinical specimens with DV200 scores below 30% requires specialized low-input targeted RNA capture panels (e.g., Illumina RNA Prep with Enrichment) or whole-transcriptome amplification chemistries. Core facility specialists possess automated liquid handling platforms that minimize hands-on pipetting variation across challenging clinical sample cohorts.
- Ultra-low input and single-cell applications: When starting biological material falls below 1 nanogram of total RNA (such as flow-sorted rare immune cell subsets, micro-dissected tissue, or single-cell suspensions), standard bulk ligation protocols fail due to sample loss and PCR bottlenecking. These samples require specialized template-switching reverse transcription chemistries (such as SMART-Seq) or microfluidic droplet partitioning platforms (such as 10x Genomics Chromium).
- Non-model organism transcriptome assembly: For research organisms lacking a published, high-quality reference genome and well-annotated gene models, read alignment cannot use standard STAR or HISAT2 pipelines. Constructing de novo transcriptomes requires specialized assembly tools (such as Trinity) coupled with transdecoder peptide identification, orthology mapping, and complex gene ontology annotation workflows supervised by computational biologists.
- Complex multifactorial and single-nuclei statistical modeling: Longitudinal time-course studies, complex cross-over treatment designs, single-nuclei pre-mRNA quantification, or multi-site clinical consortium datasets introduce nested hierarchical variance structures. Consulting a biostatistician ensures proper modeling of generalized linear mixed-effects frameworks, appropriate false discovery rate (FDR) control, and rigorous correction for technical surrogate variables.
Frequently asked questions
How much total RNA input is required for standard RNA-seq?
Standard bulk RNA-seq protocols typically require 25 ng to 1,000 ng of high-quality total RNA for oligo(dT) poly(A) mRNA enrichment, and 10 ng to 1,000 ng for ribosomal RNA depletion workflows. While newer high-sensitivity kits can generate libraries from as little as 1 ng of total RNA, using at least 100 ng to 250 ng provides robust library complexity, minimizes the required PCR amplification cycles, and significantly suppresses PCR duplication rates.
What is the minimum acceptable RIN score for RNA-seq?
For standard poly(A) enrichment RNA-seq, the minimum acceptable RNA Integrity Number (RIN) is 7.0, with a score of 8.0 or higher recommended by consortium standards to prevent 3-prime transcript coverage bias. For partially degraded samples exhibiting RIN values between 2.0 and 6.0, researchers should switch to a ribosomal RNA depletion protocol. For severely degraded FFPE samples where RIN calculation is unreliable, samples should have a DV200 score (percentage of RNA fragments greater than 200 nucleotides) of at least 30%.
Why is dUTP used in stranded RNA-seq protocols?
Deoxyuridine triphosphate (dUTP) is incorporated into the second-strand cDNA during synthesis in place of dTTP. When the library is subsequently amplified using a proofreading DNA polymerase that cannot process uracil templates, the dUTP-containing second strand is blocked from replication. Consequently, only the first-strand cDNA (which matches the antisense sequence of the original RNA molecule) is amplified, preserving the directional orientation of transcription and enabling accurate quantification of overlapping sense and antisense genes.
How many sequencing reads are needed per sample?
For standard differential gene expression profiling in mammalian transcriptomes, a sequencing depth of 20 to 30 million mapped reads per sample is standard for poly(A)-selected libraries, and 40 to 60 million mapped reads for ribosomal RNA-depleted libraries. Experiments investigating alternative splicing, novel transcript discovery, allele-specific expression, or low-abundance regulatory transcripts generally require deeper sequencing of 60 to 100 million mapped reads per sample.
What is the difference between single-end and paired-end RNA-seq?
Single-end sequencing reads each cDNA fragment from only one terminus (typically 50 to 75 bp), which is sufficient for basic gene-level quantification in well-annotated model organisms. Paired-end sequencing reads both ends of the fragment (e.g., 2 × 100 bp or 2 × 150 bp), providing spatial constraint information across the insert. Paired-end sequencing is strongly recommended for characterizing alternative splicing, detecting gene fusion events, mapping transcript isoforms, and assembling de novo transcriptomes.
Can degraded or FFPE RNA be used for RNA-seq?
Yes, degraded or archival FFPE RNA can be successfully sequenced, provided an appropriate library preparation chemistry is utilized. Oligo(dT) poly(A) selection must not be used on degraded RNA because degraded transcripts lack intact poly(A) tails or become detached from their 5-prime coding regions. Instead, researchers should use ribosomal RNA depletion kits, targeted hybridization capture panels, or 3-prime tag sequencing methods that function independently of RNA fragment length.
Sources
- Illumina, Inc. (2020). Illumina Stranded mRNA Prep Reference Guide (Document # 1000000124518 v02). https://support.illumina.com/ (accessed 2026-11-15).
- Illumina, Inc. (2020). Illumina Stranded Total RNA Prep with Ribo-Zero Plus Reference Guide (Document # 1000000124514 v02). https://support.illumina.com/ (accessed 2026-11-15).
- New England Biolabs. (2023). NEBNext Ultra II Directional RNA Library Prep Kit for Illumina Instruction Manual (NEB #E7760S/L). https://www.neb.com/ (accessed 2026-11-15).
- ENCODE Consortium. (2016). ENCODE Long RNA-Seq Pipeline Specifications and Standards. https://www.encodeproject.org/ (accessed 2026-11-15).
- Schroeder, A., Mueller, O., Stocker, S., Salowsky, R., Leenen, M., Erb, K., and Lightfoot, S. (2006). The RIN: an RNA integrity number for assigning integrity values to RNA measurements. BMC Molecular Biology, 7, 3. https://doi.org/10.1186/1471-2199-7-3 (accessed 2026-11-15).
- Dobin, A., Davis, C. A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., Chaisson, M., and Gingeras, T. R. (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics, 29(1), 15-21. https://doi.org/10.1093/bioinformatics/bts635 (accessed 2026-11-15).
- Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology, 15(12), 550. https://doi.org/10.1186/s13059-014-0550-8 (accessed 2026-11-15).
- Andrews, S. (2010). FastQC: A Quality Control Tool for High Throughput Sequence Data. Babraham Bioinformatics. https://www.bioinformatics.babraham.ac.uk/projects/fastqc/ (accessed 2026-11-15).
Data: RNA-Seq Commercial Library Kit Technical Specifications and Operating Benchmarks
Comparative technical specifications across major commercial RNA-seq library preparation kits, including total RNA input ranges, fragmentation conditions, target transcript classes, average library fragment sizes, and minimum recommended sequencing read depths compiled from manufacturer protocols and consortium guidelines.
Period covered: 2020/2026. Coverage: Global.
| Kit and Chemistry | Min Input (ng) | Max Input (ng) | Min RIN (RIN) | Enrichment Strategy | Fragmentation Protocol | Mean Insert Size (bp) | Min Recommended Depth (M reads) |
|---|---|---|---|---|---|---|---|
| Illumina Stranded mRNA Prep | 25 | 1000 | 7 | Oligo(dT) magnetic bead capture | 8 min at 94C in EPH3 buffer | 300 | 25 |
| Illumina Stranded Total RNA with Ribo-Zero Plus | 10 | 1000 | 2 | Enzymatic DNA probe rRNA depletion | 8 min at 94C in EPH3 buffer | 325 | 40 |
| NEBNext Ultra II Directional RNA (PolyA) | 10 | 1000 | 7 | Oligo(dT) magnetic bead capture | 15 min at 94C in fragmentation mix | 290 | 25 |
| NEBNext Ultra II Directional RNA (rRNA Depletion) | 10 | 1000 | 2 | RNase H targeted probe rRNA depletion | 15 min at 94C in fragmentation mix | 310 | 40 |
| Lexogen QuantSeq 3 Prime mRNA-Seq FWD | 0.5 | 500 | 1 | 3-prime poly(A) targeted oligo(dT) priming | No thermal fragmentation required | 250 | 5 |
Sources
- Illumina Stranded mRNA Prep Reference Guide (accessed 2026-11-15)
- Illumina Stranded Total RNA Prep with Ribo-Zero Plus Reference Guide (accessed 2026-11-15)
- NEBNext Ultra II Directional RNA Library Prep Kit Instruction Manual (accessed 2026-11-15)
- ENCODE Long RNA-Seq Pipeline Specifications (accessed 2026-11-15)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







