[BioC] Non linear intensities

[David martin]

Ok thanks,


Kasper Daniel Hansen wrote:
> This is well known and is (one of) the reason(s) for the difficulty with 
> analyzing microarrays.  Note that if you compute fold change, it looks a 
> lot better which is why we use fold change.  Also note that 
> normalization (hopefully) will help you a bit.
> 
> But in the end there will be genes where the fold change is not great 
> either (like 'C" in your example), and you will need to learn how to 
> live with it.
> 
> There are no great way to correct for it, since the correction typically 
> depends on the probe used to measure the signal.  If you had the money 
> and ability to generate a sample in which _all_ genes are expressed and 
> you then do a dilution experiment you might be able to estimate this.  
> There are attempts at correcting for this computationally, but they are 
> typically not very impressive.
> 
> For more details, see the 100s of papers on microarray analysis and 
> probe effects.
> 
> Kasper
> 
> On Jul 23, 2009, at 8:25 , David martin wrote:
> 
>> Hello,
>> I have a naive question. In my experiment i have computed different 
>> standard curves for different genes.
>> I have noticed that my data is not following a linear pattern (sorry 
>> i'm not statistician so don't know if this is the word).
>> For e.g
>> GENES    [1µg]    [2µg] [5µg] [10µg]
>> gene A    500    800    1800    3500
>> gene B    450    650    1700    3400
>> gene C    200    300    600    1300
>>
>> As you can see the intensities are not as I would expect (there is no 
>> linear intensity based on the concentration); Since my data shows that 
>> there is a bias i would like to correct the data and adjust according.
>> For instance, i think there might be a coeffecient, which based on the 
>> intensitiy would help to correct the intenstities. At the end if for a 
>> gene X i find an intensity of 500 i could assume that it should higher 
>> than that to have a 2 fold change and that it should probably be 
>> somewhere around 1000.
>>
>> Thanks for your help ??
>>
>> thanks for your help.
>>
>> _______________________________________________
>> Bioconductor mailing list
>> Bioconductor at stat.math.ethz.ch
>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>> Search the archives: 
>> http://news.gmane.org/gmane.science.biology.informatics.conductor
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: 
> http://news.gmane.org/gmane.science.biology.informatics.conductor
>