Dear Roumyana, this is because I rationalized the function structure, as follows: 1.) The generic function "vsn2" has methods for data objects of class AffyBatch, ExpressionSet, NChannelSet, RGList and matrix. It returns a fit object of class "vsn" with model fit parameters etc. The generic function "predict" applies a fitted model to a data object (of any of the above classes) and, in the case of AffyBatch and in most other cases, returns a data object of the same class and shape. 2.) So "justvsn" is now literally just as follows: justvsn = function(x, ...) { fit = vsn2(x, ...) predict(fit, newdata=x, useDataInFit=TRUE) } 3.) If you also want to do the probe-set summarization by RMA (an Affymetrix-specific feature that has nothing directly to do with vsn), you can call vsnrma which is a wrapper around "justvsn" and then "rma" from the affy package where the latter is called with options normalize=FALSE, background=FALSE. Bottomline: please have a look at the vsnrma function. Also, for all those of you who have endured the vsn version 1 --> version 2 transition, to all intents and purposes the reengineering of vsn is now done and I intend to make no further changes to the user interface. Best wishes Wolfgang ------------------------------------------------------------------ Wolfgang Huber EBI/EMBL Cambridge UK http://www.ebi.ac.uk/huber Roumyana Yordanova wrote: > Dear Wolfgang, > > It seems that justvsn in the last version > (http://www.bioconductor.org/packages/devel/bioc/html/vsn.html) does not > output the row.names (the corresponding probe sets names). Can this be > something related to changes in the AffyBatch object? > > Roumyana > > Roumyana Yordanova wrote: > >> Dear Wolfgang, >> Thanks for your help. For anyone that applies vsn2 to AffyBatch >> objects with a reference set please use the last version of vsn >> http://www.bioconductor.org/packages/devel/bioc/html/vsn.html >> Roumyana >> >> >> >> ------------------------------------------------------------------- ----- >> >> Subject: >> Re: vsn with a reference set >> From: >> Wolfgang Huber <huber at="" ebi.ac.uk=""> >> Date: >> Fri, 28 Mar 2008 12:05:40 +0000 >> To: >> Roumyana Yordanova <roumyana.yordanova at="" bms.com=""> >> >> To: >> Roumyana Yordanova <roumyana.yordanova at="" bms.com=""> >> >> >> 28/03/2008 04:55 Roumyana Yordanova a ?crit >> >> >>> Dear Wolfgang, >>> Thanks a lot. I got the current version and it works. >>> Roumyana >>> >>> >> >> Dear Roumyana >> >> thank you! Could you post this reply to the mailing list, so that >> everybody knows that they should upgrade their vsn if they want to work >> with reference normalisation and AffyBatch objects? >> >> >> >> > >

Dear Wolfgang, Thanks a lot for your help. Roumyana Wolfgang Huber wrote: > Dear Roumyana, > > this is because I rationalized the function structure, as follows: > > 1.) The generic function "vsn2" has methods for data objects of class > AffyBatch, ExpressionSet, NChannelSet, RGList and matrix. It returns a > fit object of class "vsn" with model fit parameters etc. The generic > function "predict" applies a fitted model to a data object (of any of > the above classes) and, in the case of AffyBatch and in most other > cases, returns a data object of the same class and shape. > > 2.) So "justvsn" is now literally just as follows: > > justvsn = function(x, ...) { > fit = vsn2(x, ...) > predict(fit, newdata=x, useDataInFit=TRUE) > } > > 3.) If you also want to do the probe-set summarization by RMA (an > Affymetrix-specific feature that has nothing directly to do with vsn), > you can call > > vsnrma > > which is a wrapper around "justvsn" and then "rma" from the affy > package where the latter is called with options normalize=FALSE, > background=FALSE. > > Bottomline: please have a look at the vsnrma function. > > Also, for all those of you who have endured the vsn version 1 --> > version 2 transition, to all intents and purposes the reengineering of > vsn is now done and I intend to make no further changes to the user > interface. > > Best wishes > Wolfgang > > ------------------------------------------------------------------ > Wolfgang Huber EBI/EMBL Cambridge UK http://www.ebi.ac.uk/huber > > Roumyana Yordanova wrote: > >> Dear Wolfgang, >> >> It seems that justvsn in the last version >> (http://www.bioconductor.org/packages/devel/bioc/html/vsn.html) does >> not output the row.names (the corresponding probe sets names). Can >> this be something related to changes in the AffyBatch object? >> >> Roumyana >> >> Roumyana Yordanova wrote: >> >>> Dear Wolfgang, >>> Thanks for your help. For anyone that applies vsn2 to AffyBatch >>> objects with a reference set please use the last version of vsn >>> http://www.bioconductor.org/packages/devel/bioc/html/vsn.html >>> Roumyana >>> >>> >>> >>> ------------------------------------------------------------------ ------ >>> >>> >>> Subject: >>> Re: vsn with a reference set >>> From: >>> Wolfgang Huber <huber at="" ebi.ac.uk=""> >>> Date: >>> Fri, 28 Mar 2008 12:05:40 +0000 >>> To: >>> Roumyana Yordanova <roumyana.yordanova at="" bms.com=""> >>> >>> To: >>> Roumyana Yordanova <roumyana.yordanova at="" bms.com=""> >>> >>> >>> 28/03/2008 04:55 Roumyana Yordanova a ?crit >>> >>> >>>> Dear Wolfgang, >>>> Thanks a lot. I got the current version and it works. >>>> Roumyana >>>> >>>> >>> >>> >>> Dear Roumyana >>> >>> thank you! Could you post this reply to the mailing list, so that >>> everybody knows that they should upgrade their vsn if they want to work >>> with reference normalisation and AffyBatch objects? >>> >>> >>> >>> >> >> >