RNA-Seq Quality Control: Essential Checks and Tools

RNA sequencing has become a standard method for measuring gene expression across entire transcriptomes, but the quality of the data depends on decisions made at every step from sample collection through computational analysis. Quality control in RNA-seq is the process of evaluating whether your sequencing data accurately represents the biological RNA present in your original samples, and it requires attention to both laboratory-derived metrics and pipeline-generated statistics. This article provides a practical framework for understanding and implementing RNA-seq quality control, with specific attention to raw read assessment, alignment metrics, gene body coverage, and sample-level checks that help you decide which data can be trusted for downstream analysis.

Why Quality Control Matters in RNA-Seq Experiments

RNA-seq is a multistep process that includes reverse transcription, amplification, fragmentation, purification, adaptor ligation, and sequencing. Improper handling at any of these stages can introduce bias or produce unusable data, and the intrinsic complexity of transcriptomes adds further challenges to accurate measurement. Comprehensive quality assessment is therefore the first and most critical step before any downstream analysis or biological interpretation can be trusted.

The consequences of skipping quality control are not abstract. Low-quality samples can distort differential expression results, create false clusters in unsupervised analyses, and lead to conclusions that do not reflect the biology of your system. In large cohort studies, where samples may be processed across multiple batches or over extended time periods, quality variation can become confounded with the biological variables you are trying to measure. Post-sequencing quality control is a crucial component of RNA-seq data generation and analysis because sample quality can be affected by sample storage, extraction, and sequencing protocols.

Quality control also serves a verification function beyond technical metrics. Studies frequently include multiple samples taken from the same individual at various time points or under different conditions, and correct assignment of those samples to each participant is evidently of great importance. Sample mix-ups are a real risk in large studies, and QC tools that can verify sample identity add an important layer of protection against this type of error.

At a Glance: RNA-Seq QC Decision Table

The table below summarizes the main quality control stages, the key metrics to examine at each stage, and the practical decisions that follow from your observations.

Raw Read Quality Control with FastQC

The first quality control step occurs before alignment, when you examine the raw sequencing reads themselves. FastQC is the most widely used tool for this purpose, and it generates a report covering multiple aspects of read quality that you should review systematically.

Per-Base Sequence Quality

The per-base quality score distribution tells you whether sequencing quality degrades across the length of the reads. Most sequencing platforms show some quality decline toward the 3' end of reads, which is expected and generally acceptable. What you are looking for is severe or early degradation that would compromise base calling accuracy across a substantial portion of each read. If quality drops below acceptable thresholds early in the read, you may need to trim reads to a shorter length before alignment.

Adapter Content and Overrepresented Sequences

Adapter contamination occurs when the sequenced fragment is shorter than the read length, causing the sequencer to read into the adapter sequence. This is common in libraries with short inserts or when fragmentation produced fragments smaller than expected. High adapter content requires trimming before alignment, because adapter sequences will not map to the genome and can create false junction alignments. Overrepresented sequences that are not adapters may indicate PCR duplication artifacts or contamination from other sources.

GC Content and Nucleotide Composition Bias

The GC content distribution of your reads should approximate the expected GC content of your organism's transcriptome. A narrow abnormal peak or a shifted distribution can indicate contamination, PCR bias, or systematic errors during library preparation. Nucleotide composition bias, where the proportion of each base varies across read positions, is often seen at the first several bases due to random hexamer priming during reverse transcription. This bias is a known intrinsic artifact of RNA-seq library preparation and is generally tolerated, but extreme patterns warrant investigation.

Duplication Levels

PCR duplication occurs when the same cDNA fragment is amplified multiple times during library preparation, producing identical reads that do not represent independent biological observations. High duplication rates reduce the effective sequencing depth and can bias expression measurements, particularly for highly expressed genes. Note that high duplication in RNA-seq can also reflect genuine biological abundance, since highly expressed transcripts naturally produce many identical reads. The interpretation of duplication metrics therefore depends on the total sequencing depth and the complexity of the library.

Alignment Metrics and Their Interpretation

After raw read QC and any necessary trimming, the next step is aligning reads to a reference genome or transcriptome. The alignment results provide a rich set of quality metrics that reflect both the sequencing run and the library preparation quality.

Total Reads and Alignment Rates

The total number of reads sequenced is the starting point for all downstream calculations. The alignment rate, or the proportion of reads that map to the reference, is one of the most informative single metrics. A low alignment rate can indicate contamination with non-target species, adapter contamination that was not removed, or problems with the reference genome itself. For most well-characterized organisms, you should expect the majority of reads to align, though the exact proportion depends on the organism, the library type, and whether ribosomal RNA depletion or poly-A enrichment was used.

Uniquely Aligned Reads and Multi-Mapping Reads

Reads that align to exactly one location in the genome are considered uniquely aligned and are the most informative for expression quantification. Reads that align to multiple locations, known as multi-mapping reads, are ambiguous because you cannot determine which copy of a repeated sequence they originated from. High proportions of multi-mapping reads can indicate that your reference genome contains many repetitive elements or that your reads are too short to be uniquely placed. The proportion of uniquely aligned reads is among the most highly correlated pipeline QC metrics with overall sample quality.

Exon, Intron, and Intergenic Reads

The genomic regions where reads align provide information about library quality and sample composition. In a typical mRNA-seq experiment using poly-A enrichment, the majority of reads should map to exons. High intronic read proportions can indicate genomic DNA contamination, incomplete splicing, or the presence of pre-mRNA in your samples. Reads mapping to intergenic regions may indicate annotation incompleteness or contamination. The expected proportions vary by organism and library preparation method, so you should establish baseline expectations for your specific experimental system.

rRNA Content

Ribosomal RNA contamination is a common problem in RNA-seq libraries. Ribosomal RNA comprises the majority of total RNA in most cells, and if depletion or enrichment steps are inefficient, rRNA reads will consume sequencing capacity without providing useful transcriptomic information. High rRNA content reduces the effective depth for mRNA quantification and indicates a problem with the library preparation protocol. The percentage of rRNA reads is one of the pipeline QC metrics most strongly correlated with overall sample quality.

Gene Body Coverage and Bias Assessment

Gene body coverage refers to the distribution of sequencing reads across the length of transcripts. An ideal RNA-seq library should produce relatively uniform coverage across the full length of each transcript, though some 3' bias is expected with poly-A enrichment protocols.

3' and 5' Bias

The 3' or 5' bias metric measures whether reads are concentrated at one end of transcripts. Excessive 3' bias can indicate RNA degradation, because degraded RNA fragments preferentially amplify from the 3' end where the poly-A tail provides the priming site. Excessive 5' bias can indicate problems with the reverse transcription or fragmentation steps. The area under the gene body coverage curve, a metric that summarizes the overall shape of coverage across gene bodies, has been identified as one of the most informative pipeline QC metrics for detecting low-quality samples.

Coverage Uniformity

Coverage uniformity across gene bodies affects the accuracy of expression quantification. If some regions of transcripts are systematically under-represented, then genes with different lengths or structures will be quantified with different biases. Tools like RNA-SeQC provide metrics for coverage continuity that help you assess whether your libraries produce even coverage. The original RNA-SeQC tool was designed to provide key measures of data quality including yield, alignment and duplication rates, GC bias, rRNA content, regions of alignment, continuity of coverage, 3' or 5' bias, and count of detectable transcripts.

GC Bias

GC bias refers to the tendency of sequencing and amplification processes to favor fragments with particular GC content. This bias can distort expression measurements, particularly for genes with extreme GC content. The GC bias metric in QC tools compares the GC content of sequenced fragments to the expected distribution from the reference transcriptome. Severe GC bias may require correction during downstream analysis or may indicate problems with the library preparation protocol.

Sample-Level Quality Checks

Beyond per-sample metrics, RNA-seq quality control should include comparisons across samples to identify outliers and verify expected relationships between samples.

Detected Gene Counts

The number of genes detected at a minimum expression threshold is a useful summary metric that reflects both sequencing depth and library quality. Samples with unusually low gene detection counts may have failed library preparation, suffered RNA degradation, or been sequenced to insufficient depth. The number of detected genes is among the pipeline QC metrics most strongly correlated with sample quality. However, the expected number of detected genes varies by tissue type, organism, and sequencing depth, so comparisons should be made within your study instead of against absolute thresholds.

Principal Component Analysis and Clustering

Principal component analysis and hierarchical clustering provide a global view of sample relationships based on their expression profiles. These analyses serve two QC purposes. First, they can reveal batch effects, where samples processed together cluster separately from samples processed at other times. Second, they can identify outliers that do not group with expected biological replicates. In a well-designed experiment, biological replicates should cluster together, and samples from different conditions should separate along the principal components that capture biological variation.

Sample Identity Verification

A critical but often overlooked QC step is verifying that each sample actually came from the individual or tissue it is labeled as. Sample mix-ups can occur during RNA extraction, library preparation, or sequencing plate setup. One approach to sample identity verification uses the highly polymorphic human leukocyte antigen genes to confirm the correct allocation of RNA-seq samples to individuals. Tools that perform HLA typing from RNA-seq data can identify wrongly assigned samples in publicly available datasets, demonstrating the importance of this type of QC for study integrity. For non-human organisms, similar approaches can use other polymorphic markers or genotype information from matched DNA samples.

Tools for RNA-Seq Quality Control

Several established tools provide the metrics described above, and your choice of tool depends on your computational environment, sample size, and specific quality concerns.

FastQC

FastQC is the standard first-line tool for raw read quality assessment. It generates an HTML report with modular assessments of per-base quality, per-sequence quality, adapter content, GC content, duplication levels, and overrepresented sequences. FastQC is fast, requires minimal computational resources, and is appropriate for any RNA-seq dataset regardless of organism or library type.

RNA-SeQC and RNA-SeQC 2

RNA-SeQC was introduced in 2012 to provide key measures of data quality including yield, alignment and duplication rates, GC bias, rRNA content, regions of alignment, continuity of coverage, 3' or 5' bias, and count of detectable transcripts. The tool was designed for multi-sample evaluation of library construction protocols, input materials, and other experimental parameters, and it enables investigators to make informed decisions about sample inclusion in downstream analysis.

RNA-SeQC 2 is an efficient reimplementation that adds multiple metrics designed to characterize sample quality across a wide range of RNA-seq protocols. The updated version was developed because RNA-seq is increasingly applied to cohorts ranging from hundreds to tens of thousands of samples, and existing tools did not readily scale to these sizes. RNA-SeQC 2 is designed for large cohorts and provides the scalability needed for modern study designs.

MultiQC

MultiQC aggregates QC reports from multiple tools and samples into a single interactive HTML report. This is particularly useful for large studies where reviewing individual FastQC or RNA-SeQC reports for hundreds of samples would be impractical. MultiQC allows you to see all samples side by side for each metric, making it easier to identify outliers and batch effects.

Integrated QC Platforms

Several platforms integrate quality control with downstream analysis steps. The Quality Control Diagnostic Renderer is software designed to simultaneously visualize a comprehensive panel of QC metrics generated by an RNA-seq pipeline and flag samples with aberrant values when compared to a reference dataset. This approach supports the conclusion that any individual QC metric is limited in its predictive value and suggests that approaches based on the integration of multiple metrics with QC thresholds are more reliable.

Single-Cell RNA-Seq Quality Control Considerations

Single-cell RNA-seq introduces additional quality control challenges beyond those encountered in bulk RNA-seq. The isolation of cells or nuclei for single-cell sequencing releases contaminating RNA through cell damage and transcript leakage, and this ambient RNA can distort biological signals. Identifying barcodes that contain high-quality cells or nuclei is a critical analytical step in the processing of single-cell RNA-seq data.

Ambient RNA Contamination

Droplet-based single-cell RNA-seq analyses assume that all acquired RNAs are endogenous to cells, but any cell-free RNAs contained within the input solution are also captured by these assays. This sequencing of cell-free RNA constitutes a background contamination that confounds the biological interpretation of single-cell transcriptomic data. This contamination from the soup of cell-free RNAs is ubiquitous, with experiment-specific variations in composition and magnitude. Tools such as SoupX can quantify the extent of the contamination and estimate background-corrected cell expression profiles that integrate with existing downstream analysis tools.

Cell and Nucleus Quality Metrics

Standard quality metrics for single-cell data include the number of unique molecular identifiers per cell, the number of genes detected per cell, and the proportion of reads mapping to mitochondrial genes. High mitochondrial read proportions typically indicate damaged cells where cytoplasmic RNA has leaked out, leaving mitochondrial transcripts over-represented. Automated methods such as valiDrops use data-adaptive thresholding on community-standard quality metrics followed by a clustering-based approach to identify barcodes with distinct biological signals. These methods can predict and flag dead cells with high accuracy, improving data quality and biological interpretation.

Doublet Detection

Doublets occur when two cells are captured in the same droplet and sequenced together, producing a mixed expression profile that does not represent any real cell type. Doublet detection is an important QC step in single-cell analysis, and several computational methods are available to identify and remove these artifacts. The SCTK-QC pipeline within the singleCellTK R package includes steps for empty droplet detection, generation of standard QC metrics, prediction of doublets, and estimation of ambient RNA.

Practical Workflow for RNA-Seq Quality Control

The following workflow provides a structured approach to implementing quality control in your RNA-seq analysis. The exact order and emphasis of steps will depend on your experimental design and the specific questions you are asking.

Step 1: Document Laboratory Metrics

Before examining sequencing data, record the laboratory-derived metrics for each sample. These include RNA integrity numbers from Bioanalyzer or TapeStation runs, RNA concentration measurements, and library preparation notes. While laboratory metrics are important for identifying problems early, note that experimental QC metrics derived from the lab may not be significantly correlated with pipeline QC metrics in some datasets. This means that a sample with acceptable lab metrics can still produce poor sequencing data, and vice versa.

Step 2: Run FastQC on Raw Reads

Run FastQC on all raw FASTQ files before any trimming or filtering. Review the per-base quality, adapter content, GC content, and duplication metrics for each sample. Flag samples with severe quality degradation, high adapter content, or abnormal GC distributions for further investigation.

Step 3: Perform Adapter Trimming and Quality Filtering

Based on the FastQC results, trim adapter sequences and filter low-quality reads using a tool such as Trimmomatic, cutadapt, or fastp. The trimming parameters should be based on the specific adapter sequences used in your library preparation and the quality profile of your reads. After trimming, run FastQC again on the trimmed reads to confirm that the problems have been addressed.

Step 4: Align Reads and Generate Alignment Metrics

Align the trimmed reads to your reference genome or transcriptome using your preferred aligner. Then run RNA-SeQC 2 or a similar tool to generate alignment-based quality metrics. Review the total aligned reads, uniquely aligned reads, exon-intron read distribution, rRNA content, and gene body coverage metrics for each sample.

Step 5: Assess Sample-Level Patterns

Generate a count matrix and examine the number of detected genes per sample. Perform principal component analysis and hierarchical clustering to visualize sample relationships. Identify outliers that do not cluster with their expected biological group and investigate whether the cause is technical or biological.

Step 6: Verify Sample Identity

If your study includes multiple samples from the same individuals or if sample mix-up is a concern, use available tools to verify sample identity. For human studies, HLA-based QC can confirm the correct allocation of RNA-seq samples to individuals. For other organisms, use whatever polymorphic markers are available in your data.

Step 7: Make Inclusion Decisions

Based on the integrated assessment of all QC metrics, decide which samples to include in downstream analysis. The decision to exclude a sample should be documented with the specific QC metrics that triggered the exclusion. In general, any individual QC metric is limited in its predictive value, so decisions should be based on the integration of multiple metrics with QC thresholds.

Records and Documentation for Quality Control

Maintaining detailed records of your quality control process is essential for reproducibility and for defending your analysis decisions in publications or regulatory submissions.

What to Record

For each sample, record the raw read count, the number of reads passing quality filters, the alignment rate, the proportion of uniquely aligned reads, the rRNA content, the number of detected genes, and any other metrics you used for inclusion decisions. Record the versions of all software tools used, including the reference genome version and annotation version. Record the specific parameters used for trimming, alignment, and QC metric calculation.

Batch and Processing Information

Record the date of RNA extraction, the date of library preparation, the sequencing run identifier, and the flow cell lane for each sample. This information is essential for identifying batch effects and for troubleshooting quality problems that cluster by processing date or sequencing run.

Exclusion Documentation

When a sample is excluded from downstream analysis, document the specific QC metrics that triggered the exclusion and the thresholds that were violated. This documentation should be included in the methods section of any publication or report arising from the data.

Common Failure Patterns and Troubleshooting

Recognizing common failure patterns can help you diagnose quality problems quickly and decide whether a sample can be rescued or must be excluded.

Low Alignment Rate

A low alignment rate, where a substantial proportion of reads do not map to the reference genome, can have several causes. Adapter contamination that was not removed during trimming will produce reads that fail to align. Contamination with DNA or RNA from other organisms, such as bacteria or fungi, will also reduce the alignment rate. If the reference genome is incomplete or poorly annotated, some reads may fail to align because the corresponding genomic sequence is missing from the reference.

High rRNA Content

High rRNA content indicates that the ribosomal RNA depletion or poly-A enrichment step in your library preparation was inefficient. This problem is often consistent across samples prepared in the same batch, so check whether the issue correlates with a particular library preparation date or technician. If rRNA content is high but the remaining reads are of good quality, you may still be able to use the data, but the effective sequencing depth will be reduced.

Severe 3' Bias

Severe 3' bias, where reads are concentrated at the 3' ends of transcripts, typically indicates RNA degradation in the original sample. Degraded RNA produces fragmented cDNA that preferentially amplifies from the 3' end. This pattern is often associated with poor sample storage or handling. Samples with severe 3' bias may still be usable for detecting highly expressed genes but will have reduced power for detecting differential expression of lowly expressed genes.

Sample Mislabeling

Sample mislabeling can occur at any stage from tissue collection through sequencing. If a sample clusters with the wrong biological group in PCA or clustering analyses, verify the sample identity using polymorphic markers. HLA-based QC has revealed wrongly assigned samples from publicly available datasets, demonstrating that this problem occurs in real studies.

Batch Effects

Batch effects appear as clustering of samples by processing date, sequencing run, or technician instead of by biological condition. Some batch effects can be corrected computationally during downstream analysis, but severe batch effects may require samples to be excluded or the experiment to be repeated. The best approach is to prevent batch effects through careful experimental design, such as randomizing sample processing across batches and including technical replicates.

Limitations of Quality Control Metrics

Quality control metrics are essential tools, but they have important limitations that you should understand when interpreting your results.

Individual Metrics Have Limited Predictive Value

Research on RNA-seq quality control has shown that any individual QC metric is limited in its predictive value. No single metric can reliably identify all low-quality samples, and the most informative metrics vary between datasets. The recommended approach is to integrate multiple metrics and use QC thresholds that are informed by the distribution of metrics across your entire dataset instead of by absolute cutoffs.

Laboratory Metrics May Not Predict Data Quality

Laboratory-derived metrics such as RNA integrity numbers may not be significantly correlated with pipeline QC metrics in some datasets. This means that samples with acceptable lab metrics can still produce poor sequencing data. The reverse is also possible, where samples with concerning lab metrics produce acceptable sequencing data. You should therefore not rely solely on laboratory metrics to make inclusion decisions.

Reference Dependence

Many QC metrics depend on the quality and completeness of the reference genome and annotation. A poorly annotated genome will produce high intergenic read proportions and low exon read proportions that do not reflect library quality. Similarly, a reference genome that is missing sequences will produce lower alignment rates. When comparing QC metrics across studies, ensure that the same reference and annotation versions were used.

Thresholds Are Dataset-Specific

The appropriate thresholds for QC metrics vary by organism, tissue type, library preparation method, and sequencing platform. Thresholds that work for one dataset may be inappropriate for another. The most defensible approach is to examine the distribution of each metric across your samples and identify outliers relative to that distribution, instead of applying fixed cutoffs from the literature.

Safety and Regulatory Context for RNA-Seq Data

RNA-seq data generated from human subjects or clinical samples is subject to regulatory requirements that extend to the quality control process.

Data Sharing Policies

If your RNA-seq data will be deposited in public repositories, you must comply with applicable data sharing policies. The NIH Genomic Data Sharing Policy sets expectations for the sharing of genomic data generated from NIH-funded research, including requirements for data quality, documentation, and controlled access for sensitive data. Review the requirements that apply to your funding source and study type before depositing data.

Data Repositories

The NCBI provides data resources for the deposition and retrieval of sequencing data, including the Sequence Read Archive and the Gene Expression Omnibus. These repositories have specific format requirements and quality expectations for submitted data. Ensuring that your QC process is documented and that your data meets repository standards will facilitate deposition and reuse.

FAIR Data Principles

The FAIR Guiding Principles describe the characteristics that data should have to be findable, accessible, interoperable, and reusable. Quality control documentation is an important component of making data reusable, because it provides the information needed for other researchers to assess whether your data are suitable for their purposes. Documenting your QC process and making that documentation available with your data supports the FAIR principles.

Professional Escalation Criteria

Knowing when to escalate a quality control problem to a supervisor, collaborator, or specialist is important for maintaining data integrity and avoiding wasted effort on unusable data.

When to Consult a Bioinformatics Specialist

If you encounter alignment rates that are dramatically lower than expected for your organism and library type, or if QC metrics show patterns that you cannot explain based on your laboratory records, consult a bioinformatics specialist. Similarly, if you are working with a non-model organism and the QC metrics suggest problems with the reference genome instead of the samples, specialist input may be needed to determine whether a different reference or annotation is required.

When to Consult a Laboratory Specialist

If QC metrics indicate consistent problems across samples prepared in the same batch, such as high rRNA content or high adapter contamination, consult the laboratory personnel responsible for library preparation. The problem may be a protocol deviation, a reagent failure, or an equipment malfunction that can be corrected for future samples.

When to Consider Resequencing

If a sample fails multiple QC metrics and the failure pattern indicates a problem that cannot be corrected computationally, such as severe RNA degradation or sample contamination, resequencing may be the most efficient path forward. The decision to resequence should consider the cost of sequencing, the importance of the sample to the study, and the likelihood that a new library preparation would produce better data.

When to Exclude a Sample

Sample exclusion is appropriate when a sample fails multiple independent QC metrics and the failure cannot be attributed to a correctable technical issue. The decision to exclude should be made in consultation with the study team and should be documented with the specific QC evidence. Excluding a sample is preferable to including low-quality data that could distort downstream results.

Frequently Asked Questions

What is the RIN value and why does it matter for RNA-seq?

The RNA integrity number is a measure of RNA quality that is calculated from the electrophoretic trace of an RNA sample. It reflects the degree of RNA degradation, with higher values indicating more intact RNA. RNA integrity is important because degraded RNA produces biased sequencing data with reduced coverage of the 5' ends of transcripts. However, laboratory-derived metrics such as RIN may not be significantly correlated with pipeline QC metrics in some datasets, so a good RIN does not guarantee good sequencing data and a poor RIN does not always predict unusable data.

What does gene body coverage tell you about library quality?

Gene body coverage describes how sequencing reads are distributed across the length of transcripts. Uniform coverage across gene bodies indicates that the library preparation produced cDNA fragments that represent the full length of transcripts. Excessive 3' bias suggests RNA degradation, while other non-uniform patterns can indicate problems with fragmentation or reverse transcription. The area under the gene body coverage curve is one of the pipeline QC metrics most strongly correlated with sample quality.

How do I decide whether to exclude a low-quality sample?

Sample exclusion decisions should be based on the integration of multiple QC metrics instead of any single metric. Examine the distribution of each metric across all samples in your study and identify samples that are clear outliers on multiple independent metrics. Document the specific metrics and thresholds that triggered the exclusion decision. If a sample fails only one metric and the failure can be explained by a known technical issue, you may be able to include the sample with appropriate caveats.

What is the difference between uniquely aligned and multi-mapping reads?

Uniquely aligned reads map to exactly one location in the reference genome and can be confidently assigned to a specific gene. Multi-mapping reads align to multiple locations, often because they come from repetitive or duplicated genomic regions, and their gene assignment is ambiguous. High proportions of multi-mapping reads can reduce the accuracy of expression quantification for genes in repetitive regions. The proportion of uniquely aligned reads is among the pipeline QC metrics most strongly correlated with sample quality.

How can I detect sample mix-ups in my RNA-seq data?

Sample mix-ups can be detected by comparing the expression profiles of samples to expected relationships, such as samples from the same individual clustering together in PCA or hierarchical clustering analyses. For human studies, HLA-based QC tools can verify the correct allocation of RNA-seq samples to individuals by typing the highly polymorphic HLA genes. These tools have revealed wrongly assigned samples in publicly available datasets, demonstrating the importance of this verification step.

What is ambient RNA contamination in single-cell RNA-seq?

Ambient RNA contamination occurs when cell-free RNA present in the input solution is captured along with the RNA from individual cells or nuclei in droplet-based single-cell sequencing. This contamination is ubiquitous and varies in composition and magnitude between experiments. It can distort biological signals and confound interpretation. Tools such as SoupX can quantify the extent of contamination and estimate background-corrected expression profiles.

Should I use the same QC thresholds for all types of RNA-seq data?

No, QC thresholds should be adapted to your specific experimental system. The expected values for metrics such as alignment rate, exon read proportion, and detected gene count vary by organism, tissue type, library preparation method, and sequencing platform. The most defensible approach is to examine the distribution of each metric across your samples and identify outliers relative to that distribution, instead of applying fixed cutoffs from the literature.

What should I do if my QC metrics indicate problems with the reference genome?

If QC metrics suggest that the reference genome is incomplete or poorly annotated, such as high intergenic read proportions or low alignment rates that cannot be explained by sample quality, consider whether an alternative reference or annotation is available. For non-model organisms, you may need to generate a de novo transcriptome assembly or use a related species as a reference. Consult a bioinformatics specialist to determine the best approach for your specific organism and research question.

Related Bioinformatics Guides

• Alternative Splicing Analysis from RNA-Seq Data
• Single-Cell RNA-seq Clustering and Cell-Type Annotation Pipelines
• Single-Cell RNA-Seq Normalization: Batch Effect Correction and Dimension Reduction (PCA, t-SNE, UMAP)
• Single-Cell RNA-Seq Analysis Pipelines for Veterinary Immunology
• Single-cell RNA-seq Trajectory Inference and Cell Lineage Tracing

References and Further Reading

• EMBL-EBI Training. European Bioinformatics Institute.
• NCBI Data Resources. National Center for Biotechnology Information.
• Genomic Data Sharing Policy. National Institutes of Health.
• The FAIR Guiding Principles. Scientific Data.
• RNA2HLA: HLA-based quality control of RNA-seq datasets.. Briefings in bioinformatics, 2021.
• Automatic quality control of single-cell and single-nucleus RNA-seq using valiDrops.. NAR genomics and bioinformatics, 2023.
• Integration of Bulk RNA-seq Pipeline Metrics for Assessing Low-Quality Samples.. Research square, 2025.
• RNA-SeQC 2: efficient RNA-seq quality control and quantification for large cohorts.. Bioinformatics (Oxford, England), 2021.
• Integration of bulk RNA-seq pipeline metrics for assessing low-quality samples.. BMC bioinformatics, 2026.
• SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data.. GigaScience, 2020.
• RNA-SeQC: RNA-seq metrics for quality control and process optimization.. Bioinformatics (Oxford, England), 2012.
• Quality control of RNA-seq experiments.. Methods in molecular biology (Clifton, N.J.), 2015.
• DiRT v2.0: An Optimized Pipeline for Detecting Dicistronic tRNA-mRNA Transcripts in Plants. 2026.
• RAGER: A user-friendly computational platform for integrated analysis of RNA-Seq and ATAC-seq data.. 2026.
• Enhanced RNA-Seq Expression Profiling and Functional Enrichment in Non-model Organisms Using Custom Annotations. 2026.
• Single-cell RNA-seq data of wild type and fli1b mutant zebrafish embryos.. 2026.
• Metatranscriptomic Reanalysis of Alzheimer's Brains Identifies Low-Biomass Microbial Signals Including Enrichment of &lt,i&gt,Acinetobacter radioresistens&lt,/i&gt,.. 2026.
• De novo transcriptome assembly of Clitoria ternatea flowers.. 2026.
• Quality Control Metrics for Extraction-Free Targeted RNA-Seq Under a Compositional Framework. Springer Proceedings in Mathematics & statistics, 2016.
• A comprehensive assessment of RNA-seq accuracy, reproducibility and information content by the Sequencing Quality Control consortium. Nature Biotechnology, 2014.
• popsicleR: A R Package for Pre-processing and Quality Control Analysis of Single Cell RNA-seq Data.. Journal of Molecular Biology, 2022.
• Comprehensive generation, visualization, and reporting of quality control metrics for single-cell RNA sequencing data. Nature Communications, 2020.

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.