[BioC] basics of GLMFit

[Gordon K Smyth]

Dear Chris,

Steve has asked you some highly relevant questions and suggested a 
reference on glms.  I will clarify some other issues.

I assume you are actually refering to the glmFit() and glmLRT() functions 
in the edgeR package (I admit that I have myself written fitLRT instead of 
glmLRT on occasion.)  These functions fit negative binomial rather than 
binomial glms.  And they do this in a parallelized fashion to multiple 
response vectors.

The algorithm and strategy of the glmFit() function is outlined in the 
publication:

   McCarthy et al "Differential expression analysis of multifactor
   RNA-Seq experiments with respect to biological variation",
   Nucleic Acids Research 2012
   http://www.ncbi.nlm.nih.gov/pubmed/22287627

I have also updated the documentation page for these functions in the past 
few days to give better points to the lower level functions that they 
call.

However you should not expect an elementary tutorial on these algorithms, 
because they are not elementary by any stretch of the imagination, or at 
least so it seems to me.  Indeed, the whole intention of the edgeR package 
is that it should give biological researchers access to statistically 
advanced algorithms without their having to worry too much about the 
numerics under the hood.

Best wishes
Gordon

> Date: Thu, 3 May 2012 09:48:03 -0400
> From: Steve Lianoglou <mailinglist.honeypot at gmail.com>
> To: chris Jhon <cjhon217 at gmail.com>
> Cc: bioconductor at r-project.org
> Subject: Re: [BioC] basics of GLMFit
>
> Hi,
>
> On Thu, May 3, 2012 at 3:59 AM, chris Jhon <cjhon217 at gmail.com> wrote:
>> Hi All,
>>
>> I read the manual of glmFit and fitLRT functions but i would appreciate if
>> any one can explain or (send me a tutorial) about the basics of fitting
>> binomial generalized linear model to a data.
>> 
>> Thank you.
>> Regards,
>> Chris
>
> Is it the "negative binomial" part, or the "glm" part that you are
> curious about learning? Or are you curious about how one actually
> "fits" a model to data?
>
> This is a hard question to answer because there's just so much to know
> and it's unclear what you mean by "the basics" since ... well ...
> depending on your background it's not all that basic ;-)
>
> Have you read through:
> http://en.wikipedia.org/wiki/Generalized_linear_model
>
> Lots of refs there including a link to McCullagh & Nelder Textbook if
> that's what you're interested in. There's also lots of things that pop
> up when you google "fit glm" (it's not just R help pages, either ;-)
>
> -steve
>
> -- 
> Steve Lianoglou
> Graduate Student: Computational Systems Biology
> ?| Memorial Sloan-Kettering Cancer Center
> ?| Weill Medical College of Cornell University
> Contact Info: http://cbio.mskcc.org/~lianos/contact

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