[BioC] edgeR glm

Gordon K Smyth smyth at wehi.EDU.AU
Thu Apr 7 01:22:27 CEST 2011


Dear Shreyartha,

logConc is a summary measure of average concentration for each tag over 
all treatment conditions.  It is the same value regardless of what 
comparison is being made.  In the version of edgeR about to be released, 
it is computed by mglmOneGroup() and is close to 
log(mean(count/lib.size)).

Best wishes
Gordon

---------------------------------------------
Professor Gordon K Smyth,
NHMRC Senior Research Fellow,
Bioinformatics Division,
Walter and Eliza Hall Institute of Medical Research,
1G Royal Parade, Parkville, Vic 3052, Australia.
Tel: (03) 9345 2326, Fax (03) 9347 0852,
smyth at wehi.edu.au
http://www.wehi.edu.au
http://www.statsci.org/smyth

On Wed, 6 Apr 2011, Shreyartha Mukherjee wrote:

> Hi Gordon,
>
> Thanks for your help. I had another question about the lrt table (obtained
> from glmlrt).
>
> How do we interpret the logConc values for 3 genotypes, I understand for a
> pairwise comparison it should be (logA + logB)/2.
> Is it (logA + logB+ logC)/3 ?
>
> Thanks,
> Shreyartha
>
>
>
>
>
> On Tue, Apr 5, 2011 at 9:11 PM, Gordon K Smyth <smyth at wehi.edu.au> wrote:
>
>> Dear Shreyartha,
>>
>> You code is actually testing for genes that are DE between f1 and p1.  By
>> default, glmLRT tests for the last coefficient in the linear model and in
>> your model this happens to represent f1-p1.  See help("glmLRT").
>>
>> What you probably want is
>>
>>  lrt <- glmLRT(d,fit,coef=c(2,3))
>>
>> This will test whether f1=p1 and p2=p1 simultaneously, i.e., whether all
>> three genotypes have the same expression.  This will find genes that are
>> differentially expressed between any of the genotypes.  This is analogous to
>> doing genewise oneway ANOVA tests for differences between the three
>> genotypes.
>>
>> Note that the above test does not guarantee that all three genotypes are
>> different.  There is actually no statistical test that can do that. Rather
>> it detects genes for which the genotypes are not all equal.
>>
>> Unfortunately, the current release version of topTags() won't work properly
>> when you test for several coefficients as above.  We'll fix this for the
>> next Bioconductor release in a couple of weeks time.  In the meantime, you
>> could use the following code:
>>
>>  o <- order(lrt$table$p.value)
>>  ntags <- 10
>>  lrt$table[o[1:ntags],]
>>
>> to see the most significant tags.
>>
>> Best wishes
>> Gordon
>>
>> ---------------------------------------------
>> Professor Gordon K Smyth,
>> NHMRC Senior Research Fellow,
>> Bioinformatics Division,
>> Walter and Eliza Hall Institute of Medical Research,
>> 1G Royal Parade, Parkville, Vic 3052, Australia.
>> Tel: (03) 9345 2326, Fax (03) 9347 0852,
>> smyth at wehi.edu.au
>> http://www.wehi.edu.au
>> http://www.statsci.org/smyth
>>
>>
>> On Tue, 5 Apr 2011, Shreyartha Mukherjee wrote:
>>
>>  Hi Gondon,Bioconductor experts,
>>>
>>> I have RNA-seq data for 3 genotypes(each genotype having 3 biological
>>> replicates) and 30,000 genes. I was trying to find out differentially
>>> expressed genes across all genotypes. Here is the code I am using.
>>>
>>> library(edgeR)
>>> y<-read.table('test_12339.txt'
>>> )
>>> d <- DGEList(y, group = rep(1:3, each = 3), lib.size = lib.sizes)
>>> d <- calcNormFactors(d)
>>> times <- rep(c("p1","p2","f1"),c(3,3,3))
>>> times <- factor(times,levels=c("p1","p2","f1"))
>>> design <- model.matrix(~factor(times))
>>> disp <- d2$tagwise.dispersion
>>> fit <- glmFit(d,design,dispersion=disp)
>>> lrt <- glmLRT(d,fit)
>>> topTags(lrt)
>>>
>>> Does this code provide me with tags differentially expressed across all
>>> conditions? (I am not looking for pairwise differential expression
>>> but across all conditions)
>>>
>>> Thanks,
>>> Shreyartha
>>>
>>
>> ______________________________________________________________________
>> The information in this email is confidential and intended solely for the
>> addressee.
>> You must not disclose, forward, print or use it without the permission of
>> the sender.
>> ______________________________________________________________________
>>
>

______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}



More information about the Bioconductor mailing list