Komplettansicht Dear all, I have received a microarray data set in standard Affymetrix CEL-format consisting of only six samples without any replicates (same organism and cell type, but different individuals and different biological conditions for each individual; the same Affymetrix GeneChip platform was used for all samples). Moreover, the data was apparently collected without any a-priori biological hypothesis. I know that it is impossible to apply standard clustering, feature selection or classification techniques in this case. However, I am wondering whether anybody is aware of a method in R to extract meaningful biological information in this case (i.e. from single-sample microarray data or from multiple samples with different biological conditions and no replicates) - or is there nothing I can do given the above limitations? Many thanks, Rainer

Rainer, I think your experiment could yield many insights. Arrays are mostly about hypothesis generation, not hypothesis testing anyway. Clustering your samples should tell you which conditions may be most similar. That should be interesting. For any pair of comparisons, you are certainly entitled to observe fold change differences of various genes, and you can take the most highly regulated genes and see whether they belong to paths. Just do some inexpensive low throughput experiments to validate anything that your array analysis uncovers. My two cents. Tom n Sep 22, 2009, at 3:16 PM, Rainer Tischler wrote: > Komplettansicht > Dear all, > > I > have received a microarray data set in standard Affymetrix CEL- format > consisting of only six samples without any replicates (same organism > and cell type, but different individuals and different biological > conditions for each individual; the same Affymetrix GeneChip platform > was used for all samples). Moreover, the data was apparently collected > without any a-priori biological hypothesis. > > I know that it is > impossible to apply standard clustering, feature selection or > classification techniques in this case. However, I am wondering > whether > anybody is aware of a method in R to extract meaningful biological > information in this case (i.e. from single-sample microarray data or > from multiple samples with different biological conditions and no > replicates) - or is there nothing I can do given the above > limitations? > > Many thanks, > Rainer > > > > > _______________________________________________ > 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

You cannot control the error rates or assign any statistical interpretation to the data. But you can still look and see what might be interesting. However, you do have to ask the investigators how much time and money they want to throw at an analysis that will have a very high false positive and false negative rate. Which depends, I suppose on the relative cost of someone's time to do the analyses and validation studies versus the cost of collecting some replicates. --Naomi At 03:16 PM 9/22/2009, Rainer Tischler wrote: >Komplettansicht >Dear all, > >I >have received a microarray data set in standard Affymetrix CEL-format >consisting of only six samples without any replicates (same organism >and cell type, but different individuals and different biological >conditions for each individual; the same Affymetrix GeneChip platform >was used for all samples). Moreover, the data was apparently collected >without any a-priori biological hypothesis. > >I know that it is >impossible to apply standard clustering, feature selection or >classification techniques in this case. However, I am wondering whether >anybody is aware of a method in R to extract meaningful biological >information in this case (i.e. from single-sample microarray data or >from multiple samples with different biological conditions and no >replicates) - or is there nothing I can do given the above limitations? > >Many thanks, >Rainer > > > > >_______________________________________________ >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

Hello all, The puma package allows you to calculate Differentially Expressed genes (function"pumaDE") utilizing single .CEL-files (no biological replicates). Be careful with interpreting the resulting data, because it does not replace the statistically more sound inclusion of biological replicates! http://www.bioconductor.org/packages/bioc/html/puma.html Regards, Dr. Philip de Groot Ph.D. Bioinformatics Researcher Wageningen University / TIFN Nutrigenomics Consortium Nutrition, Metabolism & Genomics Group Division of Human Nutrition PO Box 8129, 6700 EV Wageningen Visiting Address: Erfelijkheidsleer: De Valk, Building 304 Dreijenweg 2, 6703 HA Wageningen Room: 0052a T: +31-317-485786 F: +31-317-483342 E-mail: Philip.deGroot at wur.nl <mailto:philip.degroot at="" wur.nl=""> Internet: http://www.nutrigenomicsconsortium.nl <http: www.nutrigenomicsconsortium.nl=""/> http://humannutrition.wur.nl <http: humannutrition.wur.nl=""/> https://madmax.bioinformatics.nl <https: madmax.bioinformatics.nl=""/> ________________________________ Van: Naomi Altman [mailto:naomi at stat.psu.edu] Verzonden: wo 23-9-2009 5:25 Aan: Rainer Tischler; bioconductor at stat.math.ethz.ch Onderwerp: Re: [BioC] microarray analysis in R without replicates You cannot control the error rates or assign any statistical interpretation to the data. But you can still look and see what might be interesting. However, you do have to ask the investigators how much time and money they want to throw at an analysis that will have a very high false positive and false negative rate. Which depends, I suppose on the relative cost of someone's time to do the analyses and validation studies versus the cost of collecting some replicates. --Naomi At 03:16 PM 9/22/2009, Rainer Tischler wrote: >Komplettansicht >Dear all, > >I >have received a microarray data set in standard Affymetrix CEL-format >consisting of only six samples without any replicates (same organism >and cell type, but different individuals and different biological >conditions for each individual; the same Affymetrix GeneChip platform >was used for all samples). Moreover, the data was apparently collected >without any a-priori biological hypothesis. > >I know that it is >impossible to apply standard clustering, feature selection or >classification techniques in this case. However, I am wondering whether >anybody is aware of a method in R to extract meaningful biological >information in this case (i.e. from single-sample microarray data or >from multiple samples with different biological conditions and no >replicates) - or is there nothing I can do given the above limitations? > >Many thanks, >Rainer > > > > >_______________________________________________ >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