Dev data warning:

You are currently visiting the dev server for the IMPC. This site is classed as unstable but is displaying our test data. To access the live data from the IMPC please visit

Help & Documentation

Statistical significance

In the IMPC statistical pipeline, a genotype is called significant when it is less than 1 x 10-4 (P < 0.0001).

Sample size

As a high throughput project, the IMPC mutants’ sample size is relatively low. The target number of knockout animals being processed is normally 14 (7 females and 7 males). This is the lowest number that would consume the least amount of resources while achieving the goal of detecting phenotypic abnormalities in a strain. At times, practical issues might limit the number of animals it is possible to test, such as viability issues or the difficulty in administering a test.

The number of animals phenotyped per sex, per genotype, can be viewed on Gene pages (table view of the “All measurements” tab) and Chart pages (one for each allele and parameter), and in data files available through the FTP site or the API.

Control group

The control group is used to form parameter baselines of wild type (WT) mice and, as they are collected continuously, it is normally much larger than the mutant cohort size. The IMPC statistical pipeline tries to identify the significant  genotype effect by comparing mutants with carefully selected controls that are as homogeneous as possible to each other and which were phenotyped as close in time as possible to the mutants.

Sex effect

Since sex has been shown to be an important factor in genotype-phenotype associations, the IMPC statistical pipeline takes this factor into account by including a sex effect in the different statistical frameworks that are listed above.

Body weight effect

Body weight (BWT) is another important factor taken into account in the IMPC statistical analysis. Many phenotypes have been shown to be correlated with BWT. However, BWT is not always available for the IMPC parameters. In those cases the statistical methods are adjusted to exclude the BWT effect. In general, body weight is measured several times a week and the weighing that is closest to the date of the experiment is used as the BWT.

PhenStat User Guide 1,750 KB How to Guide – Installing PhenStat 367 KB

PhenStat is available as a Bioconductor package

Complete PhenStat user’s guide

More on statistical procedures

The IMPC Newsletter

Get highlights of the most important data releases, news and events, delivered straight to your email inbox

Subscribe to newsletter