[BioC] genotyping RLMM probe level model?

[Justin Borevitz]

Thanks Benilton! Terry sent me a copy a while back and just realizing how
good this is for 2 allele 2 strand setup (for most SNPs this is plenty).
Have you thought about allele specific expression (RNA), certainly a strand
bias with low and saturating signal + length issues? Let me know if you are
interested..

-----
Justin Borevitz
http://naturalsystems.org/lab 


-----Original Message-----
From: Benilton Carvalho [mailto:bcarvalh at jhsph.edu] 
Sent: Thursday, September 21, 2006 12:35 AM
To: Justin Borevitz
Cc: 'Ben Bolstad'; Bioconductor; 'Scott Smemo'
Subject: Re: [BioC] genotyping RLMM probe level model?

you also might want to take a look at CRLMM:

http://www.bepress.com/jhubiostat/paper111/

b

On Sep 20, 2006, at 11:25 PM, Justin Borevitz wrote:

> We'll we may give it a shot but there's always plenty of room for  
> help..
> thanks
>
> -----
> Justin Borevitz
> http://naturalsystems.org/lab
>
>
> -----Original Message-----
> From: Ben Bolstad [mailto:bolstad at stat.berkeley.edu]
> Sent: Wednesday, September 20, 2006 10:22 PM
> To: Justin Borevitz
> Cc: 'Scott Smemo'
> Subject: RE: genotyping RLMM probe level model?
>
> Oh, ok. Now I see where you are going. As far as I am aware no one is
> doing that specifically. At one point I was thinking of expanding  
> my PLM
> work in that direction, but other things got in the way and it is very
> very unlikely I'll ever get to it.
>
> Best,
>
> Ben
>
>
> On Wed, 2006-09-20 at 22:12 -0500, Justin Borevitz wrote:
>> Sorry, yes RLMM is like RMA to summarize probes prior to analysis eg
>> Classify().  But what about analysis at the probe level? I'm  
>> developing
> some
>> SNP/tiling arrays, and we're adding 4 probes (2 alleles, 2  
>> strands) for
> each
>> SNP and others for expression.  To score allele specific  
>> expression we'd
>> look at the ratio (say of the two alleles), and we can do this  
>> with both
>> strands.  With many samples each strand can predict a genotype,  
>> then the
> two
>> statistics combined rather than combine and create 1 statistic.
>>
>> Thoughts? Someone must be doing this..
>> Justin
>>
>
> _______________________________________________
> 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