RNA-Seq vs DNA-Seq: Key Differences and Applications

DNA sequencing and RNA sequencing are two distinct approaches to reading the molecular information in biological samples. DNA sequencing reveals the genetic code present in a sample, while RNA sequencing reveals which parts of that code are actively being transcribed into RNA molecules at the time of collection. This distinction matters for researchers and clinicians because the genome is largely stable throughout life, whereas the transcriptome changes constantly in response to development, environment, disease, and treatment. Choosing between RNA-seq and DNA-seq depends on the biological question being asked, the type of variation being investigated, and the practical constraints of sample availability and cost.

What Each Method Measures

DNA Sequencing Measures the Blueprint

DNA sequencing determines the order of nucleotides in a DNA molecule. This approach captures the inherited and acquired genetic variants present in a sample, including single nucleotide variants, insertions, deletions, copy number changes, and structural rearrangements. The genome serves as a stable reference point because it remains largely unchanged across cell types and over time, apart from somatic mutations that accumulate during cell division.

DNA-based next-generation sequencing is the standard approach for detecting mutations in cancer genes such as KIT and PDGFRA in gastrointestinal stromal tumors. A large DNA-sequenced cohort of 579 GIST cases found mutations in 83.2% of cases, with 403 KIT mutations and 79 PDGFRA mutations, and secondary KIT mutations in 6.0% of cases. Mutation location correlated with tumor location and patient age, and secondary mutations were more frequent in non-gastric tumors. This demonstrates how DNA sequencing provides the foundational information about which genetic alterations are present in a tumor.

Whole-exome sequencing focuses on the protein-coding regions of the genome. In a study of rejecting kidney transplants, whole-exome sequencing of donor and recipient DNA was used to identify single nucleotide variants that could distinguish cells from each participant. These DNA-sequence comparisons provided the reference variants needed to trace the origin of individual cells in the transplant.

RNA Sequencing Measures the Activity

RNA sequencing captures the transcriptome, which is the complete set of RNA molecules present in a sample at a given moment. This includes messenger RNA, non-coding RNA, and other RNA species. The transcriptome reflects which genes are actively being expressed, the relative abundance of different transcripts, and the splicing patterns that produce different protein isoforms from the same gene.

RNA-seq provides a dynamic view of cellular activity. In a study of early rheumatoid arthritis, comprehensive RNA sequencing of synovial tissue and matched peripheral blood from treatment-naive patients identified transcriptional subgroups linked to three distinct pathotypes: a fibroblastic pauci-immune pathotype, a macrophage-rich diffuse-myeloid pathotype, and a lympho-myeloid pathotype characterized by infiltration of lymphocytes and myeloid cells. These transcriptional subgroups suggested divergent pathogenic pathways and correlated with clinical response to initial drug therapy, while plasma cell genes identified a poor prognosis subgroup with progressive structural damage.

Single-cell RNA sequencing adds another dimension by measuring gene expression in individual cells instead of in bulk tissue. In a study of the unfolded protein response in the mouse intestinal epithelium, single-cell RNA sequencing defined dynamic transcriptional states associated with adaptive versus terminal responses to endoplasmic reticulum stress. This approach revealed the transcription factor QRICH1 as a key effector of the PERK-eIF2α axis that controls a transcriptional program associated with translation and secretory networks.

Core Differences in Experimental Design

Sample Requirements and Stability

DNA is relatively stable and can be extracted from fresh, frozen, or archived samples. RNA is more labile because RNA molecules degrade quickly after sample collection due to the ubiquitous presence of RNases. Sample handling for RNA-seq requires immediate stabilization, often through snap-freezing or the use of RNA-preserving reagents. This practical difference affects study design, particularly in clinical settings where sample collection and processing logistics vary.

What the Data Represent

DNA-seq data represent the genetic potential of a sample. The results are largely binary in the sense that a variant is either present or absent, though variant allele frequency provides information about the proportion of cells carrying the variant. RNA-seq data represent the functional state of a sample. Expression levels are continuous measurements that reflect the number of transcripts from each gene, and these levels can vary dramatically across conditions.

The distinction between genotype and expression is important for interpreting biological mechanisms. In a study of mismatch repair-deficient endometrial cancer, mutational MMRd tumors had higher mutation burden than epigenetic MMRd tumors, but within each category, mutation burden did not correlate with immune checkpoint blockade response. Longitudinal single-cell RNA-seq of circulating immune cells revealed contrasting modes of antitumor immunity, with effector CD8+ T cells correlating with regression of mutational MMRd tumors and activated CD16+ NK cells associated with ICB-responsive epigenetic MMRd tumors. This study shows how DNA information alone was insufficient to predict response, while RNA information about immune cell states provided mechanistic insight.

Bioinformatics Analysis Differences

DNA-seq analysis typically involves aligning reads to a reference genome, calling variants, and annotating those variants. Key steps include quality trimming, read alignment, duplicate removal, variant calling, and variant filtering. The analysis focuses on identifying differences between the sample and the reference genome.

RNA-seq analysis involves aligning reads to a reference genome or transcriptome, quantifying gene or transcript abundance, and performing differential expression analysis. Additional analyses include splicing analysis, isoform quantification, and gene set enrichment. The analysis must account for transcript length and sequencing depth when calculating expression levels.

RNA-seq data can also be used to predict the functional consequences of DNA variants. An empirical method developed through analysis of 5,145 cryptic-donors versus 86,963 decoy-donors predicted cryptic-donor activation with 87% sensitivity and 95% specificity. The most frequently recurring natural mis-splicing events at each exon-intron junction, summarized over 40,233 RNA-sequencing samples, predicted with accuracy which cryptic-donor would be activated in rare disease. This demonstrates how population-level RNA-seq data can inform the interpretation of DNA variants that affect splicing.

At a Glance

When to Use DNA Sequencing

Germline Variant Detection

DNA sequencing is the appropriate choice when the question concerns inherited genetic variation. This includes identifying disease-causing variants in Mendelian disorders, determining carrier status, and establishing genetic relationships. The stability of DNA makes it suitable for samples collected under varied conditions.

Somatic Mutation Profiling in Cancer

Tumor DNA sequencing identifies the mutations that drive cancer development and progression. This information guides targeted therapy selection and provides prognostic information. In gastrointestinal stromal tumors, DNA-based NGS is the standard for detecting KIT and PDGFRA mutations that guide targeted therapy. The large cohort study of 579 GIST cases demonstrated that mutation site correlated with tumor location and patient age, and secondary mutations were more frequent in non-gastric tumors.

Circulating Tumor DNA Detection

DNA sequencing of cell-free DNA from blood samples enables noninvasive detection and monitoring of tumors. In the CheckMate 915 study of 1,844 patients with resected stage III/IV melanoma receiving adjuvant immunotherapy, tumor-informed circulating tumor DNA was evaluated at postresection baseline and on-treatment. ctDNA positivity at baseline, seen in 16.2% of patients, and on-treatment was associated with higher risk of recurrence than ctDNA negativity with a hazard ratio of 1.97 and 95% confidence interval of 1.57 to 2.46. The test showed high specificity of 87% and modest sensitivity of 39%. ctDNA status combined with tumor mutational burden status and interferon gamma RNA signature score was more predictive of survival than ctDNA alone.

Transplant Chimerism Assessment

DNA sequencing can distinguish cells from different individuals in a transplant setting. Whole-exome sequencing of donor and recipient DNA identified single nucleotide variants that were then used to trace cell origin in single-cell RNA-seq data from kidney transplant biopsy cores. This approach distinguished recipient versus donor origin for all 81,139 cells examined and revealed that donor macrophages can persist for years post-transplantation.

When to Use RNA Sequencing

Gene Expression Profiling

RNA-seq is the method of choice when the question concerns which genes are active and at what levels. This applies to comparing disease states, identifying biomarkers, and understanding biological pathways. The rheumatoid arthritis study used RNA-seq to identify transcriptional subgroups in synovium linked to three distinct pathotypes, demonstrating how expression profiling can classify disease subtypes with different clinical outcomes.

Cell Type Identification and State Characterization

Single-cell RNA-seq identifies cell types based on their transcriptional profiles and characterizes the functional state of those cells. In the kidney transplant study, single-cell RNA-seq distinguished recipient versus donor origin for immune cells and revealed that recipient macrophages displayed inflammatory activation while donor macrophages demonstrated antigen presentation and complement signaling. Recipient-origin T cells expressed cytotoxic and proinflammatory genes consistent with an effector cell phenotype, while donor-origin T cells appeared quiescent and expressed oxidative phosphorylation genes.

Splicing and Isoform Analysis

RNA-seq captures information about alternative splicing that is not available from DNA sequencing alone. The cryptic-donor prediction study used population-level RNA-seq data from 40,233 samples to predict which cryptic-donors would be activated by splicing variants in patient DNA. This approach assists pathology consideration of possible consequences of a variant for the encoded protein and informs RNA diagnostic testing strategies.

Functional Validation of DNA Variants

RNA-seq can confirm whether DNA variants have functional consequences at the transcript level. In the GIST study, RNA-seq was performed on 24 cases previously analyzed by DNA-based NGS, including 16 KIT mutants, 5 PDGFRA mutants, and 3 wild-type cases. RNA-seq was successful in 21 cases, identifying 13 of 14 KIT mutations and all PDGFRA mutations, including single nucleotide variants, insertions, and in-frame deletions of 3 to 27 base pairs. One complex 45 base pair deletion-insertion was not detected, though low-level evidence below 10% of reads was present. This demonstrates that RNA-seq can detect clinically relevant mutations but may miss complex rearrangements.

Host-Pathogen Interaction Studies

Dual RNA-seq can simultaneously capture transcripts from a host and a pathogen. In a study of Flavobacterium psychrophilum, the causative agent of Bacterial Cold-Water disease in salmonids, RNA-seq identified 2,190 transcripts expressed in the whole cell and 2,046 transcripts in outer membrane vesicles. Of these, 168 transcripts were uniquely identified in OMVs, 312 transcripts were expressed only in the whole cell, and 1,878 transcripts were shared. A cell wall-associated hydrolase gene was the most highly expressed gene in OMVs and among the top upregulated transcripts in susceptible fish, and the sequence was conserved in 51 different strains of the pathogen.

Practical Workflow Considerations

Sample Collection and Processing

The first decision point is sample handling. For DNA-seq, samples can be stored frozen or processed from formalin-fixed paraffin-embedded tissue. For RNA-seq, samples must be stabilized quickly to prevent RNA degradation. This affects study design in clinical settings where sample collection occurs across multiple sites or over extended periods.

Library Preparation Choices

DNA-seq library preparation typically involves fragmentation, end repair, adapter ligation, and amplification. RNA-seq library preparation requires additional steps to convert RNA to complementary DNA, and may include ribosomal RNA depletion or poly-A selection depending on the RNA species of interest. The choice of library preparation method affects which transcripts are captured and the quantitative accuracy of the results.

Sequencing Depth and Coverage

DNA-seq depth requirements depend on the application. Germline variant detection typically requires lower depth, while somatic mutation detection in heterogeneous tumor samples requires higher depth to detect variants present in a minority of cells. RNA-seq depth requirements depend on the dynamic range of expression levels being measured and whether the goal is to detect low-abundance transcripts.

Quality Control Metrics

For DNA-seq, key quality metrics include sequencing depth, coverage uniformity, and variant call quality. For RNA-seq, additional metrics include mapping rate, gene body coverage, and the proportion of reads mapping to exonic versus intronic regions. These metrics help identify problematic samples and batch effects.

Data Analysis Pipelines

DNA-seq analysis pipelines typically include alignment to a reference genome, variant calling, and variant annotation. RNA-seq analysis pipelines include alignment or pseudo-alignment, transcript quantification, and differential expression analysis. The choice of analysis tools affects results, and pipelines should be validated and version-controlled for reproducibility.

Integrating DNA and RNA Sequencing

Complementary Information

DNA and RNA sequencing provide complementary information that together gives a more complete picture of biological systems. DNA sequencing identifies the genetic variants present, while RNA sequencing reveals which variants are expressed and how gene expression changes across conditions. This integration is particularly valuable in cancer research, where both the genetic alterations and their functional consequences matter.

Multi-Omics Approaches

Several studies have integrated DNA and RNA sequencing with other molecular measurements. In autosomal dominant polycystic kidney disease research, methyl-CpG binding domain protein-enriched genome sequencing, DNMT1 chromatin immunoprecipitation sequencing, and RNA-sequencing analysis identified two novel DNMT1 targets, PTPRM and PTPN22, that function as mediators of DNMT1 and the phosphorylation and activation of PKD-associated signaling pathways including ERK, mTOR, and STAT3. Whole-genome bisulfite sequencing in kidneys of patients with ADPKD versus normal individuals found that methylation of epigenetic clock-associated genes was dysregulated, supporting that epigenetic age is accelerated in the kidneys of patients with ADPKD.

In acute myeloid leukemia research, CIRCLE-seq and RNA-seq were used to characterize extrachromosomal circular DNA in 12 AML patients and 4 healthy controls. AML cells showed significantly increased eccDNA counts and gene involvement versus healthy controls, with distinct size peaks at 202 and 368 base pairs. Integrative analysis identified 570 genes upregulated at both the eccDNA and mRNA levels, including myeloid leukemia-related genes such as FLT3, RUNX1, and CD33. High expression of these genes correlated with poor outcomes in AML.

Single-Cell Multi-Omics

Single-cell technologies now allow simultaneous measurement of multiple molecular modalities from the same cell. A comprehensive benchmarking study compared CITE-seq, which measures RNA and cell surface proteins, with DOGMA-seq, which measures RNA, chromatin accessibility, and cell surface proteins in the same cell. The study found that DOGMA-seq with optimized digitonin permeabilization and its ATAC library provided more information, although its mRNA and cell surface protein libraries had slightly inferior quality compared to CITE-seq. This tradeoff between information content and data quality is an important consideration when designing single-cell multi-omics experiments.

Epigenomic Approaches

Epigenomic profiling can be integrated with DNA and RNA sequencing to understand gene regulation. In a study of small cell transformation in EGFR mutant lung adenocarcinoma, chromatin immunoprecipitation sequencing profiled histone modifications H3K27ac, H3K4me3, and H3K27me3, along with methylated DNA immunoprecipitation sequencing, assay for transposase-accessible chromatin sequencing, and RNA sequencing on 26 patient-derived xenograft tumors. Analysis of 126 epigenomic libraries revealed widespread epigenomic reprogramming between LUAD and tSCLC, with large numbers of differential H3K27ac sites, DNA methylation sites, and chromatin accessibility sites. A multianalyte cell-free DNA-based classifier integrating three epigenomic features discriminated between EGFRm LUAD versus tSCLC with an area under the receiver operating characteristic curve of 0.94.

Common Failure Patterns and Troubleshooting

RNA Degradation

RNA degradation is the most common cause of failed RNA-seq experiments. Symptoms include low mapping rates, poor gene body coverage, and 3-prime bias. Prevention requires immediate sample stabilization, proper storage conditions, and quality assessment before library preparation. If degradation is detected, the experiment may need to be repeated with fresh samples or the analysis approach adjusted.

Batch Effects

Batch effects arise from processing samples in different groups or at different times. These technical artifacts can obscure biological signals or create false associations. Mitigation strategies include randomized sample processing, inclusion of technical replicates, and statistical correction methods. Documentation of processing conditions is essential for identifying and correcting batch effects.

Low Mapping Rates

Low mapping rates in RNA-seq can result from sample contamination, adapter contamination, or poor reference genome quality. In DNA-seq, low mapping rates may indicate sample mix-ups or contamination from other species. Quality control metrics should be reviewed for every sample, and problematic samples should be flagged for further investigation.

Variant Calling Errors

DNA-seq variant calling can produce false positives from sequencing errors, alignment artifacts, or PCR duplicates. False negatives can result from low coverage or allele dropout. Validation of variants through orthogonal methods or visual inspection of alignments is recommended for clinically significant findings.

Expression Quantification Issues

RNA-seq expression quantification can be affected by transcript length bias, GC content bias, and multi-mapping reads. These biases can be addressed through appropriate normalization methods and the use of transcript-level quantification tools. The choice of reference transcriptome and annotation version affects results and should be documented.

Limitations and Interpretation Boundaries

DNA Sequencing Limitations

DNA sequencing cannot reveal which genes are active in a sample. A variant may be present in the genome but not expressed in the tissue of interest. DNA sequencing also cannot capture dynamic changes in gene expression that occur in response to environmental stimuli, treatment, or disease progression.

RNA Sequencing Limitations

RNA sequencing is sensitive to sample quality and processing conditions. The transcriptome is context-dependent, so results reflect the state of the sample at the time of collection. RNA-seq may not detect all genetic variants, particularly those in poorly expressed genes or those that trigger nonsense-mediated decay. The GIST study demonstrated this limitation when RNA-seq failed to detect one complex 45 base pair deletion-insertion, though low-level evidence below 10% of reads was present.

Interpretation Challenges

Both methods generate large datasets that require careful statistical analysis. Multiple testing corrections are necessary when analyzing thousands of genes or variants. Correlation does not establish causation, and functional validation is often needed to confirm biological significance. The clinical relevance of findings depends on the context and requires interpretation by qualified professionals.

Data Sharing and Reproducibility

Genomic data sharing is subject to policies that balance scientific openness with participant privacy. The NIH Genomic Data Sharing Policy provides the framework for sharing genomic data generated through NIH-funded research. The FAIR Guiding Principles describe the characteristics that data resources should have to support findability, accessibility, interoperability, and reusability. Researchers should be aware of these requirements when planning studies and depositing data.

Professional Escalation Criteria

When to Seek Specialized Consultation

Several situations warrant consultation with specialized professionals. If RNA-seq results show unexpected patterns that cannot be explained by the experimental design, consultation with a bioinformatics specialist may identify technical artifacts or analysis errors. If DNA-seq identifies variants of uncertain significance in a clinical context, consultation with a molecular pathologist or genetic counselor is appropriate.

When to Repeat Experiments

Experiments should be repeated when quality control metrics fall below established thresholds. This includes samples with low sequencing depth, poor mapping rates, or evidence of degradation. Repeating experiments with independent biological replicates strengthens the reliability of findings.

When to Use Alternative Methods

If RNA-seq fails to detect a variant that is expected based on DNA-seq results, alternative approaches may be needed. This could include targeted sequencing of specific genomic regions or protein-level validation. The choice of method depends on the biological question and the limitations of each approach.

Frequently Asked Questions

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

DNA sequencing reads the genetic code present in a sample, revealing variants that are stable over time. RNA sequencing reads the transcripts that are actively expressed, revealing which genes are turned on and at what levels. DNA-seq answers questions about what genetic changes are present, while RNA-seq answers questions about what the cells are doing.

Can RNA-seq replace DNA-seq for mutation detection?

RNA-seq can detect many mutations that are expressed in the transcriptome, but it cannot replace DNA-seq for all applications. The GIST study found that RNA-seq identified 13 of 14 KIT mutations and all PDGFRA mutations, but missed one complex 45 base pair deletion-insertion. DNA-seq remains the standard for comprehensive mutation detection because it captures variants regardless of expression level.

Why is RNA less stable than DNA?

RNA molecules are single-stranded and more chemically reactive than the double-stranded DNA molecule. RNA is also rapidly degraded by RNase enzymes that are present in cells and the environment. This instability requires special handling procedures for RNA-seq samples, including immediate stabilization and cold-chain processing.

What does single-cell RNA-seq add over bulk RNA-seq?

Single-cell RNA-seq measures gene expression in individual cells instead of averaging across a population. This reveals cellular heterogeneity that is hidden in bulk measurements. The kidney transplant study used single-cell RNA-seq to distinguish recipient from donor immune cells and showed that recipient and donor macrophages had distinct transcriptional profiles.

How do DNA and RNA sequencing complement each other in cancer research?

DNA sequencing identifies the mutations present in a tumor, while RNA sequencing reveals which genes are expressed and how the tumor is behaving. The endometrial cancer study showed that DNA information about mismatch repair status was insufficient to predict immunotherapy response, while RNA information about immune cell states provided mechanistic insight into contrasting modes of antitumor immunity.

What is circulating tumor DNA and why is it measured with DNA sequencing?

Circulating tumor DNA is cell-free DNA released by tumor cells into the blood. DNA sequencing of ctDNA enables noninvasive detection and monitoring of tumors. The CheckMate 915 study showed that ctDNA positivity at baseline was associated with higher risk of recurrence in melanoma patients receiving adjuvant immunotherapy, with a hazard ratio of 1.97.

What quality checks are needed before analyzing RNA-seq data?

Key quality checks include assessing RNA integrity before library preparation, reviewing sequencing depth and mapping rates after sequencing, and examining gene body coverage to detect degradation or bias. Samples that fail quality thresholds should be flagged or excluded from downstream analysis.

How should researchers choose between DNA-seq and RNA-seq for their study?

The choice depends on the biological question. If the question concerns genetic variants, inherited mutations, or somatic mutations, DNA-seq is appropriate. If the question concerns gene expression, cell state, splicing, or pathway activity, RNA-seq is appropriate. Studies that need both types of information may require integrating both methods.

Related Bioinformatics Guides

• Master Guide: Single-Cell RNA Sequencing Bioinformatics Workflows
• ChIP-Seq Bioinformatics Workflows
• RNA Seq Fastq Example: Structural Analysis and Computational Methodologies in Bioinformatics
• Single-Cell ATAC-Seq Bioinformatics
• Alternative Splicing Analysis from RNA-Seq Data

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.
• Molecular Portraits of Early Rheumatoid Arthritis Identify Clinical and Treatment Response Phenotypes.. Cell reports, 2019.
• Distinct Mechanisms of Mismatch-Repair Deficiency Delineate Two Modes of Response to Anti-PD-1 Immunotherapy in Endometrial Carcinoma.. Cancer discovery, 2023.
• Detecting Small Cell Transformation in Patients with Advanced EGFR Mutant Lung Adenocarcinoma through Epigenomic cfDNA Profiling.. Clinical cancer research : an official journal of the American Association for Cancer Research, 2024.
• Harnessing Expressed Single Nucleotide Variation and Single Cell RNA Sequencing To Define Immune Cell Chimerism in the Rejecting Kidney Transplant.. Journal of the American Society of Nephrology : JASN, 2020.
• Pretreatment and on-treatment ctDNA and tissue biomarkers predict recurrence in patients with stage IIIB-D/IV melanoma treated with adjuvant immunotherapy: CheckMate 915.. Journal for immunotherapy of cancer, 2025.
• Comprehensive benchmarking of CITE-seq versus DOGMA-seq single cell multimodal omics.. Genome biology, 2022.
• DNA methyltransferase 1 (DNMT1) promotes cyst growth and epigenetic age acceleration in autosomal dominant polycystic kidney disease.. Kidney international, 2024.
• QRICH1 dictates the outcome of ER stress through transcriptional control of proteostasis.. Science (New York, N.Y.), 2021.
• Integrated Multi-Omics Analysis Reveals an HCMV-Associated Late-Gene Signature Associated with Poor Survival in Pediatric Group 3 Medulloblastoma. 2026.
• Tracing the stemness and malignant transition in a heritable colorectal cancer Lynch Syndrome by single-cell RNA-seq analysis
• RNA-Based and DNA-Based Next-Generation Sequencing of KIT and PDGFRA Mutations in Gastrointestinal Stromal Tumors: Analytical Performance and Epidemiologic Insights.. 2026.
• Integrated CIRCLE-seq with RNA-seq to decipher the quantity, localization, and functional features of eccDNA in AML. Frontiers in Oncology, 2025.
• Dual RNA-Seq of Flavobacterium psychrophilum and Its Outer Membrane Vesicles Distinguishes Genes Associated with Susceptibility to Bacterial Cold-Water Disease in Rainbow Trout (Oncorhynchus mykiss). Pathogens, 2023.
• Empirical prediction of variant-activated cryptic splice donors using population-based RNA-Seq data. Nature Communications, 2022.
• Engineering Evaluation by Gene Categorization of Microarray and RNA-seq Data with eegc package. 2016.
• Can-Seq: a PCR and DNA sequencing strategy for identifying new alleles of known and candidate genes. Plant Methods, 2020.
• KRAS-retroviral fusion transcripts and gene amplification in arsenic-transformed, human prostate CAsE-PE cancer cells. Toxicology and Applied Pharmacology, 2020.
• Exomes of ductal luminal breast cancer patients from southwest colombia: Gene mutational profile and related expression alterations. Biomolecules, 2020.
• Acquired mutations and transcriptional remodeling in long-term estrogen-deprived locoregional breast cancer recurrences. Breast Cancer Research, 2021.

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