Abstract 2350: Comparison of normalization methods for RT-qPCR microRNA expression in cancer datasets
Bibliographic record
Abstract
Abstract Real-time quantitative PCR (RT-qPCR) is a widely used method for quantifying microRNA (miRNA) expression, but its accuracy depends on appropriate normalization. RT-qPCR is susceptible to technical variation from several sources, including sample collection and storage, miRNA quantity/quality, and extraction and amplification efficiencies. Normalization corrects for these factors, theoretically leaving behind only biological variation. With growing interest in miRNAs as biomarkers for the diagnosis and prognosis of various cancers, analyses must be accurate and consistent. Primarily, the comparative Ct method is used, which compares the expression of miRNAs to a reference value. Unfortunately, there is a lack of consistency on what to use as a reference. A standard recommendation is the average of multiple, stable miRNAs referred to as endogenous controls (ECs). Currently, there are no known universally stable miRNAs. Therefore, ECs should be selected on a per-disease and per-sample type basis. Multiple algorithms exist to identify stable ECs, but few studies indicate what, if any, steps were taken to implement them correctly. Also, many studies have used non-ideal reference values such as small nuclear RNAs or a single miRNA. This lack of agreement on protocol often leads to difficulty comparing studies and may produce incorrect results. This work addresses these inconsistencies by making recommendations for the normalization of RT-qPCR miRNA expression. We compared 2 widely used methods for EC selection, NormFinder and GeNorm, and to show how normalization influences results we also used an unstable miRNA. Only miRNAs with expression in all samples were considered as potential ECs. GeNorm is sensitive to correlated genes. Therefore, for any pair or trio of correlated miRNAs 1 was kept. NormFinder performs optimally with 5-10 candidate genes; therefore, we tested the 10 miRNAs with the lowest coefficients of variation (CV). The top 3 stable miRNAs were selected from each algorithm to serve as ECs, and we used their average expression to normalize each sample. For the single miRNA, samples were scaled based on its value. We applied these methods to 3 independent datasets: First, plasma samples from a study on canine osteosarcoma (OSA), including pre-amputation (n=45), post-amputation (n=27), and healthy controls (n=21). Second, tissue samples from a study on canine OSA, including primary tumour (n=42) and lung metastases (n=12). Third, serum samples from a study on canine lymphoma included B-cell (n=24), T-cell (n=16), and healthy controls (n=14). For datasets 1 and 3, NormFinder provided a better reduction in gene-specific CV and a more favourable cumulative distribution of the CV. For dataset 2, there was less initial variability, but NormFinder still had slightly better results. Both algorithms had pros and cons, but NormFinder consistently provides a better reduction in sample variation across datasets. Citation Format: Heather Treleaven, Latasha Ludwig, Alicia Viloria-Petit, R. Darren Wood, Ayesha Ali, Geoffrey A. Wood. Comparison of normalization methods for RT-qPCR microRNA expression in cancer datasets [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2350.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.048 | 0.096 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.004 |
| Bibliometrics | 0.003 | 0.005 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.009 | 0.007 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".