If you want to test if there is a significant circadian rhythm in your synchronised cells vs. your unsynchonised cells you can use the CircWave software for co-sinor analysis with a 24h wave: http://www.euclock.org/modules.php?name=Content&pa=showpage&pid=20.

Hope this works out for you.

Cheers,

Andrew

Hi Andrew, sorry for my late response. First of all thanks a lot for your suggestion.

I could not find any paper where CircWave is described... so I have couple of further questions:

1) what the difference between p value and r2 parameters for the cosinor? In other words if I have p

Sorry for question #1, current wave fit= p value =0.008 and R2 = 0.42. Thanks!

Hello Danielle,

The P value for the co-sinor revelas whether the 95% confidence interval for the amplitude of the fitted wave includes zero - if it does (P>0.05) then the fit is non-significant, if P

Thanks a lot ! Now it is clear!

Best!

Daniele

Dear Danielle,

for more information about statistical procedures in chronobiologic data analyses, you should learn this paper: Procedures for numerical analysis of circadian rhythms (by Refinetti, Germaine and Halberg in Biological Rhythm Research). Go also to www.circadian.org and use a free software cosinor if neccessary!

Thanks Mamame that's also a good suggestion!

you can use the softwares in the following link:

http://www.circadian.org/softwar.html

and of course - plot your data and look at it.

also - a simple 1-way ANOVA is a common tool to detect the effect of time on response rate, the cosinor analysis is to determine whether your data wave is a cosine.

I also suggest using JTK_Cycle; Hughes ME, Hogenesch JB, Kornacker K.

J Biol Rhythms. 2010 Oct;25(5):372-80; it was developed for expression profiling but is very useful for other types of time series data.

Please, someone know some program in the RStudio  to circadian gene expression???? Thanks. 

yes, JTK-Cycle (Hughes et al., 2010).  Information about the program can be found here:  http://openwetware.org/wiki/HughesLab:JTK_Cycle

My lab and many others use this program fairly routinely for targeted gene expression (qPCR) as well as array data.

Thanks very much Karyn Esser.

I have used three tools: RAIN, empirical-JTK_Cycle and Haystack. The first runs in R env and the latters by command line. You can also run Haystack on line (google Haystack mockler lab). All the three work with curve-fitting to detect rhythms and Haystack looks more simple. However, you do not need to run the statistics by yourself because the tools will do this to you. You just need to input data and parameters.

Rain -> http://journals.sagepub.com/doi/full/10.1177/0748730414553029

Haystack -> http://haystack.mocklerlab.org/

e_JTK ->http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1004094

I try to detect a circadian rhythm with cosinor function using R. I got these results (see picture). There are p-values for acrophase and amplitude but is there any information about significant rhythmicity? As shown in the picture, p < 0.001 for amplitude and p > 0.6814 for acrophase (if X=1). Does it mean there is no circadian rhythm?

I am a bit confused with this package since there is no information on the tested hypothesis...

Thank you in advance for your help.

Sebastien Baekelandt : the null hypothesis is that amplitude of oscillations equals 0. The p-value of oscillation amplitudes thus corresponds to the significance of rhythmicity.

Best,

Miha

If you prefer Python - I have recently implemented a Python package for cosinor-based rhytmometry, which extends the functionalities of existing R packages. You can install the package using pip (pip install CosinorPy) or download it from https://github.com/mmoskon/CosinorPy

The documentation with some examples is available at the repository mentioned above.

Miha