Dear Ron, I noticed that you specified totalTest=100 which is too small for your dataset. It should be larger than the largest peak number in your datasets. Here is a list of old posts on makeVennDiagram that should help you to choose an appropriate number for totalTest. –https://stat.ethz.ch/pipermail/bioconductor/2010-June/033941.html –https://stat.ethz.ch/pipermail/bioconductor/2010-November/036540.html –http://comments.gmane.org/gmane.science.biology.informatics.conductor /32345 –https://stat.ethz.ch/pipermail/bioconductor/attachments/20100905/bb68 9e19/attachment.pl –http://permalink.gmane.org/gmane.science.biology.informatics.conducto r/29476 – http://permalink.gmane.org/gmane.science.biology.informatics.conduct or/30115 –http://permalink.gmane.org/gmane.science.biology.informatics.conducto r/35629 –http://comments.gmane.org/gmane.science.biology.informatics.conductor /34765 Best regards, Julie On 3/12/12 3:33 PM, "Ron Hart" <rhart@rci.rutgers.edu> wrote: Hi Julie, Once again, thank you for writing such a useful R package! I’m working on analyzing some data for another manuscript and I’m having trouble generating a p-value from makeVennDiagram. Enclosed is a text file with sample output. Can you help me figure out why I get NaN for the p.value? If it helps there are something like 50,000 peaks per sample. Thanks very much, Ron [[alternative HTML version deleted]]

Thanks, Julie. If you notice, the second trial in my text file uses totalTest=100000, which should be sufficient, right? Ron From: Zhu, Lihua (Julie) [mailto:Julie.Zhu@umassmed.edu] Sent: Monday, March 12, 2012 3:50 PM To: Ron Hart Cc: bioconductor@r-project.org Subject: Re: question about makeVennDiagram Dear Ron, I noticed that you specified totalTest=100 which is too small for your dataset. It should be larger than the largest peak number in your datasets. Here is a list of old posts on makeVennDiagram that should help you to choose an appropriate number for totalTest. -https://stat.ethz.ch/pipermail/bioconductor/2010-June/033941.html -https://stat.ethz.ch/pipermail/bioconductor/2010-November/036540.html -http://comments.gmane.org/gmane.science.biology.informatics.conductor /32345 -https://stat.ethz.ch/pipermail/bioconductor/attachments/20100905/bb68 9e19/a ttachment.pl -http://permalink.gmane.org/gmane.science.biology.informatics.conducto r/2947 6 - http://permalink.gmane.org/gmane.science.biology.informatics.conductor /30115 -http://permalink.gmane.org/gmane.science.biology.informatics.conducto r/3562 9 -http://comments.gmane.org/gmane.science.biology.informatics.conductor /34765 Best regards, Julie On 3/12/12 3:33 PM, "Ron Hart" <rhart@rci.rutgers.edu> wrote: Hi Julie, Once again, thank you for writing such a useful R package! I'm working on analyzing some data for another manuscript and I'm having trouble generating a p-value from makeVennDiagram. Enclosed is a text file with sample output. Can you help me figure out why I get NaN for the p.value? If it helps there are something like 50,000 peaks per sample. Thanks very much, Ron [[alternative HTML version deleted]]

Ron, For your dataset, I notice that one dataset contains 173122 peaks. Therefore, set totalTest=100000 is too small. It should be large than 173122. The totalTest number is used as space size to determine if the overlap between two datasets is more than what would be expected by chance. So it should be equal to the number of possible binding events. As an example, if you estimate that the total number of possible binding sites in the genome of interest is 250000 by motif search or other methods, then you would put totalTest=250000. Here is a nice discussion on how to set totalTest by Noah. https://stat.ethz.ch/pipermail/bioconductor/2010-November/036540.html Hope this helps! Best regards, Julie On 3/12/12 4:05 PM, "Ron Hart" <rhart at="" rci.rutgers.edu=""> wrote: > Thanks, Julie. If you notice, the second trial in my text file uses > totalTest=100000, which should be sufficient, right? > > Ron > > > From: Zhu, Lihua (Julie) [mailto:Julie.Zhu at umassmed.edu] > Sent: Monday, March 12, 2012 3:50 PM > To: Ron Hart > Cc: bioconductor at r-project.org > Subject: Re: question about makeVennDiagram > > Dear Ron, > > I noticed that you specified totalTest=100 which is too small for your > dataset. It should be larger than the largest peak number in your datasets. > > Here is a list of old posts on makeVennDiagram that should help you to choose > an appropriate number for totalTest. > > ?https://stat.ethz.ch/pipermail/bioconductor/2010-June/033941.html > ?https://stat.ethz.ch/pipermail/bioconductor/2010-November/036540.html > ?http://comments.gmane.org/gmane.science.biology.informatics.conduct or/32345 > ?https://stat.ethz.ch/pipermail/bioconductor/attachments/20100905/bb 689e19/att > achment.pl > ?http://permalink.gmane.org/gmane.science.biology.informatics.conduc tor/29476 > ? http://permalink.gmane.org/gmane.science.biology.informatics.condu ctor/30115 > ?http://permalink.gmane.org/gmane.science.biology.informatics.conduc tor/35629 > ?http://comments.gmane.org/gmane.science.biology.informatics.conduct or/34765 > > Best regards, > > Julie > > > On 3/12/12 3:33 PM, "Ron Hart" <rhart at="" rci.rutgers.edu=""> wrote: > Hi Julie, > > Once again, thank you for writing such a useful R package! I?m working on > analyzing some data for another manuscript and I?m having trouble generating a > p-value from makeVennDiagram. Enclosed is a text file with sample output. > > Can you help me figure out why I get NaN for the p.value? If it helps there > are something like 50,000 peaks per sample. > > Thanks very much, > > Ron > > >