[BioC] Agilent Mouse 8x60K array
Nathan (Nat) Goodman
ngoodman at systemsbiology.org
Wed Feb 6 15:00:57 CET 2013
Thanks for the detailed response, Gordon. limma is a great package that we use all the time. Evidently, we should be using it even more! It will take me a few days to work though your suggestions. I'll get back with any questions along the way and conclusions at the end.
Thanks again,
Nat
On Feb 5, 2013, at 4:48 PM, Gordon K Smyth wrote:
> Dear Nat,
>
> Are your arrays hybed with one dye or two? I will assume one.
>
>> Date: Mon, 4 Feb 2013 12:31:36 -0800
>> From: "Nathan (Nat) Goodman" <ngoodman at systemsbiology.org>
>> To: "James W. MacDonald" <jmacdon at uw.edu>
>> Cc: bioconductor at r-project.org
>> Subject: Re: [BioC] Agilent Mouse 8x60K array
>>
>> Hi Jim
>>
>> Everything you mentioned is good, and I agree straightforward to program up by hand. The other things I'd like to do are equally obvious and probably not too hard.
>>
>> 1) Use the negative controls to define the limit of detection.
>
> The propexpr() function in the limma package does this.
>
> The nec() and necq() functions also use the negative controls in the context of background correction and normalization, but the local background estimates provided by Agilent should be subtracted first.
>
>> 2) Use the positive controls to define the standard curve -- aka normalization -- or at least confirm that normalization worked as expected.
>
> limma also uses the controls, and not just the positive controls, in the normalization process.
>
> limma offers rich possibilities to up-weight or down-weight different types of controls in various ways, even to determine the normalization entirely from controls, but I doubt that there is any need to do this for a Mouse 8x60K array.
>
> To examine how well the normalization has worked with respect to the controls, use the plotMA() function after setting probe control status appropriately.
>
>> 3) Propagate the variance estimates from the replicated probes to downstream tests of significance.
>
> limma does this using duplicateCorrelation(). Otherwise, if the replicated probes don't fit into the duplicateCorrelation framework, then propogating the variances is essentially impossible, for the reasons explained by Jim.
>
> BTW, you asked for an Agilent equivalent of the affy package, but the affy package doesn't do (2) or (3) for Affymetrix arrays.
>
> Best wishes
> Gordon
>
>> Before I forget, I want to thank you for taking the time to engage in this conversation. I really appreciate the help.
>>
>> Best,
>> Nat
>>
>
> ______________________________________________________________________
> The information in this email is confidential and inte...{{dropped:6}}
More information about the Bioconductor
mailing list