Oneclass SAM analysis
1
0
Entering edit mode
@brooke-powell-elizabeth-1185
Last seen 7.4 years ago
Dear list, I am wondering if you would be able help me with a problem I seem to be experiencing using SAM (v2.20a). I am using it through the Excel add- in, but the question I have is not a technical computer or programming problem (to my knowledge) but one about the analysis itself. When I am performing oneclass analysis of ratio data from a microarray experiment it seems that the oneclass analysis only likes to have odd numbers of data columns where each data column is a list of ratios for one slide in an experiment. I can get analysis done for 3,5,7,9,and 11 slides of data, but not 2,4,6,8,10 or 12. I am wondering if I am missing something about oneclass analysis, but it seems to me that you would want a balanced design of an experiment with even numbers of slides. Any help would be appreciated, I have tried contacting the SAM list which is next to useless. Happy New Year to everyone, Liz Brooke-Powell Washington University Medical School, Department of Molecular Microbiology, CB#8230, 660 S. Euclid Ave St. Louis, MO, 63110
• 615 views
ADD COMMENT
0
Entering edit mode
Jianping Jin ▴ 890
@jianping-jin-1212
Last seen 7.4 years ago
Hi Elizabeth, Did you set every measurement equal to 1s (not 1s and 2s)? As I understand one class analysis tests whether the mean gene expression across all replicates differs from zero. There is no any group comparison involved. If you were trying to do groups 1 and 2 comparison you probably need a "two-class" test. hope this helps Jianping --On Thursday, January 04, 2007 3:22 PM -0600 "Brooke-Powell, Elizabeth" <etbp2 at="" borcim.wustl.edu=""> wrote: > Dear list, > > I am wondering if you would be able help me with a problem I seem to be > experiencing using SAM (v2.20a). I am using it through the Excel add-in, > but the question I have is not a technical computer or programming > problem (to my knowledge) but one about the analysis itself. > > When I am performing oneclass analysis of ratio data from a microarray > experiment it seems that the oneclass analysis only likes to have odd > numbers of data columns where each data column is a list of ratios for one > slide in an experiment. I can get analysis done for 3,5,7,9,and 11 slides > of data, but not 2,4,6,8,10 or 12. I am wondering if I am missing > something about oneclass analysis, but it seems to me that you would want > a balanced design of an experiment with even numbers of slides. > > Any help would be appreciated, I have tried contacting the SAM list which > is next to useless. > > Happy New Year to everyone, > > Liz Brooke-Powell > > Washington University Medical School, > Department of Molecular Microbiology, > CB#8230, 660 S. Euclid Ave > St. Louis, MO, 63110 > > _______________________________________________ > 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 ################################## Jianping Jin Ph.D. Bioinformatics scientist Center for Bioinformatics Room 3133 Bioinformatics building CB# 7104 University of Chapel Hill Chapel Hill, NC 27599 Phone: (919)843-6105 FAX: (919)843-3103 E-Mail: jjin at email.unc.edu
ADD COMMENT
0
Entering edit mode
Jianping, Thank you for getting back to me, I do set all of the array column headers to 1 as all of the biological comparisons are the same in the same ratio orientation. It has to do with the number of columns of data in total on the worksheet, it only likes odd numbers of arrays. Liz -----Original Message----- From: Jianping Jin [mailto:jjin@email.unc.edu] Sent: Friday, January 05, 2007 8:25 AM To: Brooke-Powell, Elizabeth; bioconductor at stat.math.ethz.ch Subject: Re: [BioC] Oneclass SAM analysis Hi Elizabeth, Did you set every measurement equal to 1s (not 1s and 2s)? As I understand one class analysis tests whether the mean gene expression across all replicates differs from zero. There is no any group comparison involved. If you were trying to do groups 1 and 2 comparison you probably need a "two-class" test. hope this helps Jianping
ADD REPLY
0
Entering edit mode
It is odd. I am not sure how you know the plug-in just took the odd numbers of columns? I would run the sample data (came along with the SAM plug-in) for one class test and see how it works. Or you may need to update the plug-in to the current version and try it out. Jianping. --On Friday, January 05, 2007 9:34 AM -0600 "Brooke-Powell, Elizabeth" <etbp2 at="" borcim.wustl.edu=""> wrote: > Jianping, > > Thank you for getting back to me, I do set all of the array column headers > to 1 as all of the biological comparisons are the same in the same ratio > orientation. It has to do with the number of columns of data in total on > the worksheet, it only likes odd numbers of arrays. > > Liz > > > -----Original Message----- > From: Jianping Jin [mailto:jjin at email.unc.edu] > Sent: Friday, January 05, 2007 8:25 AM > To: Brooke-Powell, Elizabeth; bioconductor at stat.math.ethz.ch > Subject: Re: [BioC] Oneclass SAM analysis > > Hi Elizabeth, > > Did you set every measurement equal to 1s (not 1s and 2s)? As I > understand one class analysis tests whether the mean gene expression > across all replicates differs from zero. There is no any group > comparison involved. If you were trying to do groups 1 and 2 comparison > you probably need a "two-class" test. > > hope this helps > > Jianping > > > ################################## Jianping Jin Ph.D. Bioinformatics scientist Center for Bioinformatics Room 3133 Bioinformatics building CB# 7104 University of Chapel Hill Chapel Hill, NC 27599 Phone: (919)843-6105 FAX: (919)843-3103 E-Mail: jjin at email.unc.edu
ADD REPLY
0
Entering edit mode
Actually I tested it. I had 12 slides worth of data, and tried to run SAM and it choked, so I went and tried with only 3 slides which worked and slowly increased the numbers of columns by one each time up to 12. Every time I had an even number of slides it choked and every time I had an odd number of columns; at least up to 11; it worked fine. I have what is the latest version according to the SAM website, but it was a curious problem and wondered if there was a mathematical/statistical reason as to why even numbers of columns don't work. From experimental design aspects it is better to have a balanced design with an even number of hybridizations. Liz -----Original Message----- From: Jianping Jin [mailto:jjin@email.unc.edu] Sent: Friday, January 05, 2007 11:06 AM To: Brooke-Powell, Elizabeth; bioconductor at stat.math.ethz.ch Subject: RE: [BioC] Oneclass SAM analysis It is odd. I am not sure how you know the plug-in just took the odd numbers of columns? I would run the sample data (came along with the SAM plug-in) for one class test and see how it works. Or you may need to update the plug-in to the current version and try it out. Jianping. --On Friday, January 05, 2007 9:34 AM -0600 "Brooke-Powell, Elizabeth" <etbp2 at="" borcim.wustl.edu=""> wrote: > Jianping, > > Thank you for getting back to me, I do set all of the array column headers > to 1 as all of the biological comparisons are the same in the same ratio > orientation. It has to do with the number of columns of data in total on > the worksheet, it only likes odd numbers of arrays. > > Liz > >
ADD REPLY
0
Entering edit mode
I installed the SAM v.2.23A on a machine with the Windows XP OS and have never observed what you described. A problem I ever ran into was memory overflow. I am not a statistician, but I doubt there are any statistical reasons for preference of odd number of samples in one class test. Jianping --On Friday, January 05, 2007 11:30 AM -0600 "Brooke-Powell, Elizabeth" <etbp2 at="" borcim.wustl.edu=""> wrote: > Actually I tested it. I had 12 slides worth of data, and tried to run SAM > and it choked, so I went and tried with only 3 slides which worked and > slowly increased the numbers of columns by one each time up to 12. Every > time I had an even number of slides it choked and every time I had an odd > number of columns; at least up to 11; it worked fine. > > I have what is the latest version according to the SAM website, but it > was a curious problem and wondered if there was a > mathematical/statistical reason as to why even numbers of columns don't > work. From experimental design aspects it is better to have a balanced > design with an even number of hybridizations. > > Liz > > -----Original Message----- > From: Jianping Jin [mailto:jjin at email.unc.edu] > Sent: Friday, January 05, 2007 11:06 AM > To: Brooke-Powell, Elizabeth; bioconductor at stat.math.ethz.ch > Subject: RE: [BioC] Oneclass SAM analysis > > It is odd. I am not sure how you know the plug-in just took the odd > numbers of columns? I would run the sample data (came along with the > SAM plug-in) for one class test and see how it works. Or you may need to > update the plug-in to the current version and try it out. > Jianping. > > --On Friday, January 05, 2007 9:34 AM -0600 "Brooke-Powell, Elizabeth" > <etbp2 at="" borcim.wustl.edu=""> wrote: > >> Jianping, >> >> Thank you for getting back to me, I do set all of the array column >> headers to 1 as all of the biological comparisons are the same in the >> same ratio orientation. It has to do with the number of columns of data >> in total on the worksheet, it only likes odd numbers of arrays. >> >> Liz >> >> > ################################## Jianping Jin Ph.D. Bioinformatics scientist Center for Bioinformatics Room 3133 Bioinformatics building CB# 7104 University of Chapel Hill Chapel Hill, NC 27599 Phone: (919)843-6105 FAX: (919)843-3103 E-Mail: jjin at email.unc.edu
ADD REPLY
0
Entering edit mode
A one-class test is correct if the data are the log expression ratios (M). --Naomi At 09:24 AM 1/5/2007, Jianping Jin wrote: >Hi Elizabeth, > >Did you set every measurement equal to 1s (not 1s and 2s)? As I understand >one class analysis tests whether the mean gene expression across all >replicates differs from zero. There is no any group comparison involved. If >you were trying to do groups 1 and 2 comparison you probably need a >"two-class" test. > >hope this helps > >Jianping > > >--On Thursday, January 04, 2007 3:22 PM -0600 "Brooke-Powell, Elizabeth" ><etbp2 at="" borcim.wustl.edu=""> wrote: > > > Dear list, > > > > I am wondering if you would be able help me with a problem I seem to be > > experiencing using SAM (v2.20a). I am using it through the Excel add-in, > > but the question I have is not a technical computer or programming > > problem (to my knowledge) but one about the analysis itself. > > > > When I am performing oneclass analysis of ratio data from a microarray > > experiment it seems that the oneclass analysis only likes to have odd > > numbers of data columns where each data column is a list of ratios for one > > slide in an experiment. I can get analysis done for 3,5,7,9,and 11 slides > > of data, but not 2,4,6,8,10 or 12. I am wondering if I am missing > > something about oneclass analysis, but it seems to me that you would want > > a balanced design of an experiment with even numbers of slides. > > > > Any help would be appreciated, I have tried contacting the SAM list which > > is next to useless. > > > > Happy New Year to everyone, > > > > Liz Brooke-Powell > > > > Washington University Medical School, > > Department of Molecular Microbiology, > > CB#8230, 660 S. Euclid Ave > > St. Louis, MO, 63110 > > > > _______________________________________________ > > 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 > > > >################################## >Jianping Jin Ph.D. >Bioinformatics scientist >Center for Bioinformatics >Room 3133 Bioinformatics building >CB# 7104 >University of Chapel Hill >Chapel Hill, NC 27599 >Phone: (919)843-6105 >FAX: (919)843-3103 >E-Mail: jjin at email.unc.edu > >_______________________________________________ >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 Naomi S. Altman 814-865-3791 (voice) Associate Professor Dept. of Statistics 814-863-7114 (fax) Penn State University 814-865-1348 (Statistics) University Park, PA 16802-2111
ADD REPLY

Login before adding your answer.

Traffic: 492 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6