Name: geNorm
Author: Jo Vandesompele

Question 1

Is geNorm free?

Accepted Answer

Yes. The core geNorm analysis on Clarida is completely free - no account required, no credit card. Create a free account to save your analyses and revisit them any time, or continue into the full qPCR workflow when you need it.

Question 2

Does geNorm work with my instrument's export format?

Accepted Answer

The geNorm tool accepts Cq values from most qPCR instruments (Bio-Rad CFX, Applied Biosystems QuantStudio, Roche LightCycler, etc.). Export your Cq values and paste or upload them directly.

Question 3

How many reference genes should I select?

Accepted Answer

geNorm calculates pairwise variation (V-value) to determine the optimal number. The general recommendation from the original paper is to use at least two reference genes. The V-value between consecutive normalization factors helps you decide when adding another gene no longer significantly improves normalization - typically when V falls below 0.15.

Question 4

Why is reference gene selection important?

Accepted Answer

Normalization against reference genes corrects for non-biological variation between samples - differences in RNA input, RNA integrity, reverse transcription efficiency, and overall transcriptional activity. If your reference genes are unstable, every downstream result is unreliable. The original geNorm paper showed that normalizing with a single unvalidated reference gene introduced errors of up to 3-fold in 25% of cases, and up to 6.4-fold in 10% (Vandesompele et al., 2002). The [MIQE guidelines (Bustin et al., 2009)](https://pubmed.ncbi.nlm.nih.gov/19246619/) classify justification of reference gene selection as Essential for publication.

Question 5

How do I validate my reference genes?

Accepted Answer

Select at least 8 candidate reference genes from different functional classes and test them across a representative set of samples that spans your experimental conditions. Run the panel through a stability algorithm like geNorm, which ranks candidates by M-value (lower = more stable) and calculates a V-value to determine how many reference genes you need. That is exactly what this tool does - import your Cq values and get a validated ranking in seconds.

Question 6

What is geNorm?

Accepted Answer

geNorm is an algorithm for selecting the most stable reference genes (also called housekeeping genes) from a set of candidates in a qPCR experiment. It was published by Vandesompele et al. in Genome Biology in 2002 and has since become the standard method, with [over 23,000 citations](https://scholar.google.be/scholar?cites=10339963667484587183).

Question 7

What is the geNorm M-value?

Accepted Answer

The M-value is a measure of expression stability calculated by geNorm. It represents the average pairwise variation of a gene with all other candidate reference genes. Lower M-values indicate more stable expression. Typical thresholds are M < 1.0 for heterogeneous sample panels or M < 0.5 for homogeneous samples (Hellemans et al., 2007). These are empirical guidelines - the original geNorm algorithm uses relative ranking rather than fixed cutoffs.

Question 8

Can geNorm rank down to the single most stable gene?

Accepted Answer

Yes. The original 2002 geNorm algorithm ranks candidate reference genes by stepwise exclusion of the least stable gene, but it stops one step short: the final pair of most stable genes share an identical M-value and cannot be told apart. This is inherent to how the M-value is defined - it is a measure of average pairwise expression stability, so once only two candidates remain they are each other's sole comparator and necessarily yield the same value. The original method therefore identifies the two most stable genes as a pair without ranking them against each other. The implementation here resolves that final step and ranks all candidates down to the single most stable gene. This enhancement was first introduced in [qbase+](/what-happened-to-qbase-plus) and is carried forward in Clarida's Reference Gene Finder. In practice the two top genes are both excellent reference targets, and for normalization you typically use them together (geNorm's V-value usually recommends two or more genes anyway), so the distinction rarely changes which genes you pick. It does give a complete, unambiguous ranking, which is useful when you need to report a single most stable gene or trim a candidate panel to a defined size.

Question 9

Who created geNorm?

Accepted Answer

geNorm was created by Jo Vandesompele and colleagues at Ghent University, Belgium. The algorithm was published in Genome Biology in 2002. Jo is a scientific advisor at Clarida. The implementation here is built on the original algorithm, with one improvement: it ranks all candidates down to the single most stable gene, where the original could only identify the best pair.

Question 10

Should I use multiple reference gene algorithms?

Accepted Answer

No. While several authors commonly use multiple reference gene validation algorithms, it is not necessary to use more than one if that single algorithm is reliable and robust. geNorm and NormFinder are two such algorithms. Own comparisons and the scientific literature confirm that the stability rankings from both algorithms are very similar, and when they deviate, the differences are marginal in a regime (of stability values) where it does not matter much which reference genes to use (e.g., there are many good choices).

Question 11

How does geNorm compare to NormFinder?

Accepted Answer

Both geNorm and NormFinder are excellent algorithms to find stable reference genes, but they use different mathematical approaches. geNorm calculates pairwise variation across all candidate reference genes, while [NormFinder (Andersen et al., 2004)](https://aacrjournals.org/cancerres/article/64/15/5245/511517/Normalization-of-Real-Time-Quantitative-Reverse) uses a model-based approach that accounts for inter- and intra-group variation. In general, rankings are very similar between the methods, but they may differ, particularly when samples have a distinct group structure. Some researchers use both algorithms and apply rank-aggregation analysis to produce a final list. geNorm is more widely cited (~23,000 vs ~6,500 citations) and is the default method in most qPCR guidelines.

Question 12

How does geNorm compare to BestKeeper?

Accepted Answer

geNorm and [BestKeeper (Pfaffl et al., 2004)](https://link.springer.com/article/10.1023/B:BILE.0000019559.84305.47) take fundamentally different approaches. geNorm works with expression ratios and calculates pairwise variation, while BestKeeper operates on raw Cq values using standard deviation and Pearson correlation. A key limitation of BestKeeper is that it does not correct for PCR efficiency and provides no guidance on the optimal number of reference genes (geNorm does, via the V-value). BestKeeper rankings tend to be the least concordant with other algorithms. Many researchers use both methods and compare results.

Question 13

How does geNorm compare to the comparative delta-Ct method?

Accepted Answer

The [comparative delta-Ct method (Silver et al., 2006)](https://pmc.ncbi.nlm.nih.gov/articles/PMC1609175/) is conceptually similar to geNorm - both use pairwise comparisons across candidate genes. The delta-Ct method calculates the standard deviation of Cq differences between gene pairs, while geNorm uses expression ratios and stepwise exclusion. The delta-Ct method is simpler (can be done in a spreadsheet) but does not provide guidance on how many reference genes to use. Both methods share a susceptibility to favoring co-regulated genes.

Reference gene finder - geNorm

How Clarida implements it

The method you trust, without the friction

Origin

Frequently asked questions

Roadmap

Are your reference genes reliable?