Hi,
I did outgrowth assays on spiral ganglion neurons from the inner ear with each supplemented by different concentrations of two neurotrophins. For each explant I did an image analysis and calculated the number of outgrown neurites and the median length. Additionally, I have further information like the location (turn) within the cochlea, which is summarized in the attached table. I did at least 6 explants per concentration and turn = 18 explants per concentration.
I would like to know if increasing neurotrophin concentrations correlate with neurite numbers and neurite lengths. Furthermore, I'm interested if there's an additive effect of the two neurotrophins (I also did 4 different "mixes"), their effect on the different turns and how the other variables influence the dependent variables.
So far I know that the explant size has a strong correlation with neurite numbers. As the size of the explant is a result of dissection I would say that it is a confounding variable and I would like to control for it. Neurite numbers and their length also correlate (probably some supportive effect).
I suspect, that there is some kind of (non-linear) dose-response relationship between the neurtotrophin concentration and the resulting variables. What I've learned so far, is that I probably have to do a nonlinear regression analysis, but I haven't found any model, which can in-cooperate all the information I have. Does anyone have an idea which model/software I could use?
SPSS, Matlab, OriginPro, Minitab and Prism are available.
Thanks a lot for your help!
See the link: https://www.bing.com/search?q=How+to+do+a+dose+response+analysis+in+R&qs=n&form=QBRE&sp=-1&pq=how+to+do+a+dose+response+analysis+in+r&sc=1-39&sk=&cvid=8350E576427D4E3EAE99D33A738488FC Cmon guys this is 2020
D. Booth
This link may also be valuable in finding an appropriate model:
https://www.bing.com/search?q=Finding+an+appropriate+dose+response+model&qs=n&form=QBRE&sp=-1&pq=finding+an+appropriate+dose+response+model&sc=1-42&sk=&cvid=1906E31691284506821C69A9B29EBBAA
See particularly:
Article Dose-Response Analysis Using R
If unfamiliar with R, see the link:
http://sgpwe.izt.uam.mx/files/users/uami/gma/R_for_dummies.pdf
D. Booth
PS Here' a tip: Consider what you will do with the data after collecting it as part of your experimental design process.
And a somewhat less helpful link but it may contain some useful information:
https://www.bing.com/search?q=multivariable+dose+response+analysis+with+confounding+variables&FORM=AWRE
And for an off the wall idea here's an example from Biochemical Kinetics that may be of some help:
https://www.bing.com/search?q=Michaelis+Menten+model+fitted+by+GAMS&qs=n&form=QBRE&sp=-1&pq=michaelis+menten+model+fitted+by+gams&sc=0-37&sk=&cvid=73965CB453FB4EEE839B7517B94070B7
An even better version of the last link is here:
https://www.bing.com/search?q=Michaelis+Menten+model+fitted+by+GAMS%28Generalized+Additive+models%29&qs=n&form=QBRE&sp=-1&pq=michaelis+menten+model+fitted+by+gams%28generalized+additive+mod&sc=1-62&sk=&cvid=11CF249B64FB419BB7CF150A201CBE0B
BTW I think you are ok treating your DVs as continuous but you might want to ask a more experienced person in this area than I am.
Dear Dominik Thomas Schmidbauer
I think that if your outcome variable is cell count you can use Poisson regression which is a non-linear model. I use Stata software.
Remember that outcome of poisson should be as event/person-time. Therefore, you should have same time period and same population (here is the number primary cells).
If you ran into a problem during working with stata please ask here.
Remember that I am not a statistician and it is my personal experience.
Regards!
Thanks a lot for your replies!
I will have a look at them and check if they help.
Hi
My expertise is in the planning and management of clinical trials, further to a CESAM certificate (Certificat d'Etudes Statistiques Appliquées à la Médecine).
Regards
Dear Dominik Thomas Schmidbauer
Remember that Poisson regression is "Log-linear". Although its plot is not linear, but its log is linear during analysis.
No problem. Both Poisson regression and nonlinear regression can be performed in Stata.
In addition, if you have several times for each sample you can consider panel regression (longitudinal regression) in Stata.
Regards!
- Badges
- Science topic