Question

Nimblegen and Ringo package

0

Entering edit mode

João Fadista ▴ 500

@joao-fadista-1942

Last seen 9.6 years ago

An embedded and charset-unspecified text was scrubbed... Name: not available Url: https://stat.ethz.ch/pipermail/bioconductor/attachments/20070816/ 72e41860/attachment.pl

• 1.8k views

ADD COMMENT • link updated 16.7 years ago by Todd Richmond ▴ 140 • written 16.7 years ago by João Fadista ▴ 500

score 0 · Answer 1 · 2007-08-16

0

Entering edit mode

Todd Richmond ▴ 140

@todd-richmond-367

Last seen 9.6 years ago

Hi Jo?o, NimbleGen does use the qspline method (normalize.qspline from the affy package) for normalizing CGH data. The tukey.biweight mean is used to scale ChIP-chip ratio gff files so that the data is centered on zero. We do not use qspline, or indeed any other normalization method, for ChIP-chip data, since we have not found a method that does not adversely affect downstream data quality. Since Ringo is a package for analyzing ChIP-chip data, I'm assuming that's why that's there. I'm curious, though, why you are trying to use it for CGH data, when there are a number of other packages specifically mean for CGH analysis. Todd > -----Original Message----- > From: bioconductor-bounces at stat.math.ethz.ch [mailto:bioconductor- > bounces at stat.math.ethz.ch] On Behalf Of Jo?o Fadista > Sent: Thursday, August 16, 2007 2:46 AM > To: bioconductor at stat.math.ethz.ch > Subject: [BioC] Nimblegen and Ringo package > > Dear all, > > I am using the Ringo package to analyse some demo CGH data from Nimblegen > (ftp://ftp.nimblegen.com/pub/demo_data/). > > 1 - In the NimbleScan User?s Guide version 2.3 it is written that they use > the qspline normalization method (Christopher Workman, et al. "A new non- > linear normalization method for reducing variability in dna microarray > experiments". Genome Biology 2002) which I believe is the one implemented > in the affy package, right? Despite of this, the Ringo package has a > supposed nimblegen normalization method related to the tukey.biweight > function also appearing in the affy package. > Maybe I am wrong, but why did the Ringo package named a supposed nimblegen > normalization method to a method which as far as I know is not from > nimblegen? > > 2 - Another question is related to the probeAnno.rda file which is needed > in some function implemented in the Ringo package. Since I don?t have > these kind of files I was wondering if I have the start/end positions and > the indices of the probes is there any easy way to make the probeAnno.rda > files? P.S. If it helps, In the demo data I have .pos, .ndf, .pair and > .gff files. > > Best regards > > Jo?o Fadista > Ph.d. student > > > > UNIVERSITY OF AARHUS > Faculty of Agricultural Sciences > Dept. of Genetics and Biotechnology > Blichers All? 20, P.O. BOX 50 > DK-8830 Tjele > > Phone: +45 8999 1900 > Direct: +45 8999 1900 > E-mail: Joao.Fadista at agrsci.dk <mailto:joao.fadista at="" agrsci.dk=""> > Web: www.agrsci.org <http: www.agrsci.org=""/> -- Todd Richmond NimbleGen Systems, Inc. Manager of Research Informatics One Science Court Madison, WI 53711 1-608-218-7600

ADD COMMENT • link 16.7 years ago Todd Richmond ▴ 140

0

Entering edit mode

Hi Joern and Todd, Thanks for the replies. Indeed, after the reply from Joern I found in the Nimblegen website that the the qspline is for normalizing CGH data while the tukey.biweight is for scaling ChIP-chip data. I am analysing the demo Nimblegen CGH data with Ringo package because I think is the only bioconductor package that can convert the .pair raw files from nimblegen into RGList, right? After I get the RGList then I can use the packages suitable for CGH analysis. By the way, I came up with another question. As I read through the CGH NimbleScan User?s Guide I didn?t find a file that describes the statistics of the fluorescent spots for each probe, like the GenePix software does (SD of the pixel intensity, Signal to noise ratio, a score for each spot, etc.). Can you generate that kind of files? If not, is it possible to transform a .ndf into a .gal file (http://www.m oleculardevices.com/pages/software/gn_genepix_file_formats.html#gal) in order to be able to analyse the nimblegen .tiff images in GenePix? Best regards, Jo?o Fadista -----Original Message----- From: Todd Richmond [mailto:todd@nimblegen.com] Sent: Thursday, August 16, 2007 3:53 PM To: Jo?o Fadista; bioconductor at stat.math.ethz.ch Subject: RE: [BioC] Nimblegen and Ringo package Hi Jo?o, NimbleGen does use the qspline method (normalize.qspline from the affy package) for normalizing CGH data. The tukey.biweight mean is used to scale ChIP-chip ratio gff files so that the data is centered on zero. We do not use qspline, or indeed any other normalization method, for ChIP-chip data, since we have not found a method that does not adversely affect downstream data quality. Since Ringo is a package for analyzing ChIP-chip data, I'm assuming that's why that's there. I'm curious, though, why you are trying to use it for CGH data, when there are a number of other packages specifically mean for CGH analysis. Todd > -----Original Message----- > From: bioconductor-bounces at stat.math.ethz.ch [mailto:bioconductor- > bounces at stat.math.ethz.ch] On Behalf Of Jo?o Fadista > Sent: Thursday, August 16, 2007 2:46 AM > To: bioconductor at stat.math.ethz.ch > Subject: [BioC] Nimblegen and Ringo package > > Dear all, > > I am using the Ringo package to analyse some demo CGH data from > Nimblegen (ftp://ftp.nimblegen.com/pub/demo_data/). > > 1 - In the NimbleScan User?s Guide version 2.3 it is written that they > use the qspline normalization method (Christopher Workman, et al. "A > new non- linear normalization method for reducing variability in dna > microarray experiments". Genome Biology 2002) which I believe is the > one implemented in the affy package, right? Despite of this, the Ringo > package has a supposed nimblegen normalization method related to the > tukey.biweight function also appearing in the affy package. > Maybe I am wrong, but why did the Ringo package named a supposed > nimblegen normalization method to a method which as far as I know is > not from nimblegen? > > 2 - Another question is related to the probeAnno.rda file which is > needed in some function implemented in the Ringo package. Since I > don?t have these kind of files I was wondering if I have the start/end > positions and the indices of the probes is there any easy way to make > the probeAnno.rda files? P.S. If it helps, In the demo data I have > .pos, .ndf, .pair and .gff files. > > Best regards > > Jo?o Fadista > Ph.d. student > > > > UNIVERSITY OF AARHUS > Faculty of Agricultural Sciences > Dept. of Genetics and Biotechnology > Blichers All? 20, P.O. BOX 50 > DK-8830 Tjele > > Phone: +45 8999 1900 > Direct: +45 8999 1900 > E-mail: Joao.Fadista at agrsci.dk <mailto:joao.fadista at="" agrsci.dk=""> > Web: www.agrsci.org <http: www.agrsci.org=""/> -- Todd Richmond NimbleGen Systems, Inc. Manager of Research Informatics One Science Court Madison, WI 53711 1-608-218-7600

ADD REPLY • link 16.7 years ago João Fadista ▴ 500

0

Entering edit mode

> -----Original Message----- > From: Jo?o Fadista [mailto:Joao.Fadista at agrsci.dk] > Sent: Thursday, August 16, 2007 9:49 AM > To: Todd Richmond; bioconductor at stat.math.ethz.ch; toedling at ebi.ac.uk > Subject: RE: [BioC] Nimblegen and Ringo package > > I am analysing the demo Nimblegen CGH data with Ringo package because I > think is the only bioconductor package that can convert the .pair raw > files from nimblegen into RGList, right? After I get the RGList then I can > use the packages suitable for CGH analysis. Steven McKinney posted some scripts to get NimbleGen data into snapCGH awhile ago (do a search for NimbleGen + snapCGH). snapCGH will then give you access to at least 3 different algorithms. > By the way, I came up with another question. As I read through the CGH > NimbleScan User?s Guide I didn?t find a file that describes the statistics > of the fluorescent spots for each probe, like the GenePix software does > (SD of the pixel intensity, Signal to noise ratio, a score for each spot, > etc.). Can you generate that kind of files? > NimbleScan can generate a feature report, which has the mean and SD of the pixel intensities comprising the feature, but that's about it. Our features are only 3x3 pixels, in a checkerboard arrangement, so there's not really enough real estate surrounding each feature to get a reasonable estimate of local background. > If not, is it possible to transform a .ndf into a .gal file > (http://www.moleculardevices.com/pages/software/gn_genepix_file_form ats.ht > ml#gal) in order to be able to analyse the nimblegen .tiff images in > GenePix? It's possible to use GenePix 6.0 to extract the .tif images, since it does support square features and rectangular positioning. But when I've tried that, it appears to only quantify a 2x2 pixel feature with no background information - probably due to the close packing of the features. You might be able to make a custom .gal file which would only quantify the "on" positions of the checkboard, thus letting GenePix make some background calculation. But we have no tools to convert a .ndf file to a .gal file... Regards, Todd -- Todd Richmond NimbleGen Systems, Inc. Manager of Research Informatics One Science Court Madison, WI 53711 1-608-218-7600

ADD REPLY • link 16.7 years ago Todd Richmond ▴ 140

score 0 · Answer 2 · 2007-08-16

Hello Jo?o, Jo?o Fadista wrote: > Dear all, > > I am using the Ringo package to analyse some demo CGH data from Nimblegen (ftp://ftp.nimblegen.com/pub/demo_data/). > > 1 - In the NimbleScan User?s Guide version 2.3 it is written that they use the qspline normalization method (Christopher Workman, et al. "A new non-linear normalization method for reducing variability in dna microarray experiments". Genome Biology 2002) which I believe is the one implemented in the affy package, right? Despite of this, the Ringo package has a supposed nimblegen normalization method related to the tukey.biweight function also appearing in the affy package. > Maybe I am wrong, but why did the Ringo package named a supposed nimblegen normalization method to a method which as far as I know is not from nimblegen? > I am not totally sure if it is still the case, but NimbleGen have used the tukey-biweight scaling to produce the data *.gff files from the scanned two-channel intensities. See http://www.nimblegen.com/products/chip/index.html and scroll down to the FAQ and show all topics. Various manuals mention it as well. I did not find this scaling terribly useful and did mainly include if for sake of completeness. > > 2 - Another question is related to the probeAnno.rda file which is needed in some function implemented in the Ringo package. Since I don?t have these kind of files I was wondering if I have the start/end positions and the indices of the probes is there any easy way to make the probeAnno.rda files? P.S. If it helps, In the demo data I have .pos, .ndf, .pair and .gff files. > Yes, there is. Have a look in the scripts directory of the installed Ringo package. In there you will find an R script called "makeProbeAnno.R" that shows how to produce a probeAnno environment from NimbleGen's POS files. Please get back to me, if you encounter any problems with this script. I would also suggest to download and install the current development version of Ringo (version > 1.1.8) from the Bioconductor web site at http://www.bioconductor.org/packages/2.1/bioc/html/Ringo.html since I streamlined that script considerably after the original release. Best regards, Joern -- Joern Toedling EMBL - European Bioinformatics Institute Wellcome Trust Genome Campus Hinxton, Cambridge CB10 1SD United Kingdom Phone +44(0)1223 492566 Email toedling at ebi.ac.uk