Dear Tuuli and others I included Spearman in my presentation and send an updated version. I calculated Spearman rhos myself; the different filtering criteria created minor differences with the ones you calculated. I was a bit surprised to see the high correlation between Spearman and Pearson on log transformed data. They show much more correlation than Spearman and OPS-Pearson. Peter ________________________________ From: Tuuli Lappalainen [mailto:Tuuli.Lappalainen@unige.ch] Sent: Tuesday, August 07, 2012 6:03 PM To: Hoen, P.A.C. 't (HG) Cc: geuvadis_rna_analysis(a)lists.crg.es Subject: Re: [Geuvadis_rna_analysis] OPS transformation (Micha) followed by Pearson correlation Hi Peter and others, This looks good. In my analyses, I've calculated the matrices of pairwise correlations using only exons/genes/transcripts (whatever the quantification unit is) that are >0 in both samples. I don't know if this matters much, but we should agree on Thursday what's the set of exons/etc that we use to calculate the correlations. I've done some comparisons of OPS+Pearson vs Spearman, and here are two scatterplots - they give very consistent results. There's one plot of all pairwise correlations, and another one of d-statistics (=median of one sample's correlations against all others). In these plots I've dropped the two expression level outliers, NA18861.4.M_120208_5 and NA19144.4.M_120208_2. Additionally, in case you don't already have these yourself, I just uploaded the correlation matrices of all 667x667 samples, both Spearman and OPS+Pearson correlations in /upload/geuvadis/wp4_rnaseq/main_project/analysis_data/qc/correlation_matrices/ . best, Tuuli Tuuli Lappalainen, PhD Department of Genetic Medicine and Development University of Geneva Medical School CMU / Rue Michel-Servet 1 1211 Geneva 4 Switzerland Tel. +41-(0)22-3795550 tuuli.lappalainen@unige.ch<mailto:tuuli.lappalainen@unige.ch> On 8/7/12 10:26 AM, P.A.C._t_Hoen@lumc.nl<mailto:P.A.C._t_Hoen@lumc.nl> wrote: Dear all We are all aware of the pitfalls of Pearson correlations on largely skewed data. Micha mentioned in Barcelona the OPS (optimal power space transformation package) that he developed. It worked very well to determine correlations between samples, I believe. Attached some results. Best Peter Dr. Peter A.C. 't Hoen Center for Human and Clinical Genetics Leiden University Medical Center Postal zone S4-P PO Box 9600 2300 RC Leiden The Netherlands phone: +31-71-5269421 fax: +31-71-5268285 e-mail: p.a.c.hoen@lumc.nl<mailto:p.a.c.hoen@lumc.nl> _______________________________________________ Geuvadis_rna_analysis mailing list Geuvadis_rna_analysis@lists.crg.es<mailto:Geuvadis_rna_analysis@lists.crg.es> http://davinci.crg.es/mailman/listinfo/geuvadis_rna_analysis