Hi everyone, I have used Nimblegen oligo arrays for ChIP-chip experiments. I now have my results of which genomic regions are getting enriched in my experiment. I have calculated some values like GC-content of all the sequences on the array and I would like to test if the distributions of these values remain unchanged in the group of enriched sequences. My intuition was the wilcoxon (Mann-Whitney) test but after discussing with a colleague I really don t know if this is the best approach. Additionally, if Wilcox.test is the correct approach, do I need to pre-process the data? ie should I just put in the function the raw values of every probe or should I first group them and put the frequencies for each value in the function? I tried both approaches and they produce opposite results. I am worried that when I feed the raw values, there is a bias to accept the alternative hypothesis because of the huge difference in the actual number of values in the two groups. Could anyone point me to the correct statistical direction? Finaly,what does the "location shift" that the Wilcox.test output refers to mean? Is it the difference of medians? Thank you Niki

"location shift" is the difference in medians. --Naomi At 02:00 PM 8/17/2007, r.athanasiadou wrote: >Hi everyone, >I have used Nimblegen oligo arrays for ChIP-chip experiments. > >I now have my results of which genomic regions are getting enriched in my >experiment. > >I have calculated some values like GC-content of all the sequences on the >array and I would like to test if the distributions of these values remain >unchanged in the group of enriched sequences. > >My intuition was the wilcoxon (Mann-Whitney) test but after discussing with >a colleague I really don t know if this is the best approach. > >Additionally, if Wilcox.test is the correct approach, do I need to >pre-process the data? ie should I just put in the function the raw values of >every probe or should I first group them and put the frequencies for each >value in the function? I tried both approaches and they produce opposite >results. > >I am worried that when I feed the raw values, there is a bias to accept the >alternative hypothesis because of the huge difference in the actual number >of values in the two groups. > > >Could anyone point me to the correct statistical direction? > >Finaly,what does the "location shift" that the Wilcox.test output refers to >mean? Is it the difference of medians? > >Thank you >Niki > >_______________________________________________ >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