how to obtain the curve fitting for all proteins
1
0
Entering edit mode
@bioinformatics-10931
Last seen 2.8 years ago
United States

 

Actually I know that they set the algorithm in a way that 

Only proteins with fold changes bigger than [fcCutoff * (1 - fcTolerance) or smaller than 1/(fcCutoff * (1 - fcTolerance))] will be used for curve fitting. Additionally, the proteins fulfilling the fcCutoff criterion without tolerance will be marked in the output column meets_FC_requirement.

 

I tried to set the fcCutoff = 1,fcTolerance = 0

because I thought 1*1 =1 for bigger situation and 1/1 will be 1 so I will set them somehow to a number that I will be able to plot the curve for all proteins. 

and I got the following error 

Fitting 3663 individual dose response curves to 3663 proteins.
Error in fitDRCurve(protID = unique(fcSplit[[i]]$id), expName = unique(fcSplit[[i]]$experiment),  : 
  task 557 failed - "missing value where TRUE/FALSE needed"

 

I want to obtain the curve fitting for all proteins and not only those that met the above criteria , how can I do it?

I also would like to know the fcCutoff and fcTolerance they use or how to select a optimal one. 

What I take their data example and use analyzeTPPCCR which automatically do all the steps.

Many thanks in advance again

 

 

tpp • 1.2k views
ADD COMMENT
1
Entering edit mode
dchilds ▴ 40
@dorotheechilds-7752
Last seen 22 months ago
Germany

Dear Mohammad,

yes, if you want to obtain dose response curves and pEC50 values for all proteins, it is possible to set fcCutoff = 1. Then, all proteins for which the fold change increases at the highest drug concentration will be marked as 'stabilized' and all others as 'destabilized'. The transformation then takes place as shown in Box 3 of Franken et al. (2015). Nature Protocols, 10(10), 1567–1593. Please be aware that many of the proteins now marked as stabilized or destabilized may not be truly affected by the drug and that the apparent increase or decrease is due to random variation. It is advisable to carefully inspect the fitted curves and apply appropriate filters on the goodness-of-fit (measured here by the R^2 value) before further processing.

The default thresholds used are fcCutoff = 1.5 (according to Box 3 in the paper) and fcTolerance = 0.1. Please see also ?analyzeTPPCCR for all the default values and further details.

The bug you reported has been fixed. It originated from a division by 0 during the transformation with fcCutoff = 1.

ADD COMMENT
0
Entering edit mode

@dorothee.childs this is totally true. what I need to know is how to have the pEC50 for all proteins ? instead just putting a threshold and making it as stabilised and destabilised, I also need to have the values for unaffected. By the way, can you please give the script that you plotted the figure in box 3 of Franken et al. (2015). Nature Protocols10(10), 1567–1593 ? it is totally impossible for me to reproduce the same plot as you reported in that protocol 

ADD REPLY

Login before adding your answer.

Traffic: 526 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6