Dear List, I have a basic statistical question regarding a Microarray data set. However, I am not an statistician. I was told the following problem can be modeled in R/ S_PLUS. I have a data set with 3 replicates for the control and 3 replicates for experimental samples for a particular experiment. I performed a t.test and adjusted the p. values for the false discovery rate. Choosing a p value < 0.05, I didn't see any significant difference between the control and the experimental samples. My question is how can I take the existing data and sort of 'simulate" (add data points) so that the adjusted p values become significant (< 0.05). The data is follows: C.1 C.2 C.3 E.1 E.2 E.3 9.33 8.97 9.03 10.12 9.84 9.65 7.35 6.98 7.14 8.36 9.30 9.02 ...... ....... ........ ........ ....... ..... ....... ....... ........ ........ ....... ...... I thank you for any suggestions/advice...............any insight into this would be a great learning experience for me. Sohail Khan Scientific Programmer COLD SPRING HARBOR LABORATORY Genome Research Center 500 Sunnyside Boulevard Woodbury, NY 11797 (516)422-4076

Well, we peformed microarray (Affy) data analysis for an investigator using limma, 3 reps control and 3 reps experimental, and did not see any significant genes after "fdr" adjustment. From previous studies they were expecting cetrain expression changes. So, the question they asked is. How many replicates it would take to see any p values below 0.05. Is it possible to do sample size calculations based on the existing data? Thanks. -Sohail -----Original Message----- From: Adaikalavan Ramasamy [mailto:ramasamy@cancer.org.uk] Sent: Wednesday, April 12, 2006 7:04 PM To: Khan, Sohail Cc: Bioconductor at stat.math.ethz.ch Subject: Re: [BioC] Predicting number of replicates Err, what are your objectives here ? Is it to do sample size calculations or simply wanting to have data to publish it. On Wed, 2006-04-12 at 17:14 -0400, Khan, Sohail wrote: > Dear List, > > I have a basic statistical question regarding a Microarray data set. However, I am not an statistician. I was told the following problem can be modeled in R/ S_PLUS. > > I have a data set with 3 replicates for the control and 3 replicates for experimental samples for a particular experiment. I performed a t.test and adjusted the p. values for the false discovery rate. Choosing a p value < 0.05, I didn't see any significant difference between the control and the experimental samples. My question is how can I take the existing data and sort of 'simulate" (add data points) so that the adjusted p values become significant (< 0.05). The data is follows: > > C.1 C.2 C.3 E.1 E.2 E.3 > 9.33 8.97 9.03 10.12 9.84 9.65 > 7.35 6.98 7.14 8.36 9.30 9.02 > ...... ....... ........ ........ ....... ..... > ....... ....... ........ ........ ....... ...... > > I thank you for any suggestions/advice...............any insight into this would be a great learning experience for me. > > > Sohail Khan > Scientific Programmer > COLD SPRING HARBOR LABORATORY > Genome Research Center > 500 Sunnyside Boulevard > Woodbury, NY 11797 > (516)422-4076 > > _______________________________________________ > 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 >

Err, what are your objectives here ? Is it to do sample size calculations or simply wanting to have data to publish it. On Wed, 2006-04-12 at 17:14 -0400, Khan, Sohail wrote: > Dear List, > > I have a basic statistical question regarding a Microarray data set. However, I am not an statistician. I was told the following problem can be modeled in R/ S_PLUS. > > I have a data set with 3 replicates for the control and 3 replicates for experimental samples for a particular experiment. I performed a t.test and adjusted the p. values for the false discovery rate. Choosing a p value < 0.05, I didn't see any significant difference between the control and the experimental samples. My question is how can I take the existing data and sort of 'simulate" (add data points) so that the adjusted p values become significant (< 0.05). The data is follows: > > C.1 C.2 C.3 E.1 E.2 E.3 > 9.33 8.97 9.03 10.12 9.84 9.65 > 7.35 6.98 7.14 8.36 9.30 9.02 > ...... ....... ........ ........ ....... ..... > ....... ....... ........ ........ ....... ...... > > I thank you for any suggestions/advice...............any insight into this would be a great learning experience for me. > > > Sohail Khan > Scientific Programmer > COLD SPRING HARBOR LABORATORY > Genome Research Center > 500 Sunnyside Boulevard > Woodbury, NY 11797 > (516)422-4076 > > _______________________________________________ > 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 >