Re: [Geuvadis_rna_analysis] RNAseq TC - minutes

Hello all, We had a teleconference today to discuss the feedback that we got from Nature. The next call will be in one week, *Thursday 31st at 4pm CET. * *A**ction item*s: - Everyone: by next week's call, please think/discuss within your group what kind of future companion projects you are working or are planning to launch based on Geuvadis data/samples. - all who have added content to the wiki: please go through the material that you have there: delete outdated content, and if needed: add material and reorganize. Some missing content is tolerable, incorrect information is not. A brief summary of today's discussion: - Manolis and myself have sent Nature's editor Magdalena Skipper a rebuttal letter to address the points raised by the reviewers, and the plan is to discuss with her soon. Let's hope that she will give us the chance to revise the paper after all - we think that there's a decent chance. - On the call there was discussion of some of the most important questions that the reviewers raised and the potential improvements based on that (regardless of whether we'll resubmit to Nature or to another journal). Depending on what happens with Nature, I'll coordinate the details of the revisions later with the respective people. - Figure 1c raised questions but it's a pretty strong analysis already. Include the splicing manuscript in the next submission will show that we've worked on these questions in further depth - data access: from the Geuvadis browser we'll provide downloads of the full master files as well as regions. The Ensembl broswer where the bams can be visualized doesn't have a slicing functionality and it's well beyond our resources to start building that. Tuuli will provide Natalja with a list of the final master files of quantifications etc. that we'll want to share once the paper is out. - Jean is doing some additional analysis to estimate how robust the transcript variation statistics are to lowly expressed gene as well as lowly expressed transcripts of highly expressed genes. - In addition to transcript ratio QTLs, we could also do polyadenylation QTLs - Pedro has quantified these already, and it might add to the novelty of the paper and address the question of variation in transcript structure. - Regarding the robustness of the analysis, Peter had a good idea of taking the old eQTLs from April (different mapper, normalization and eQTL method, analysis by population) to show that the results are highly overlapping (I believe and hope that they are). We could also e.g. quantify 5 samples with Cufflinks and show that flux is similar but better, or map to personalized genomes, but we need to think very carefully how deep we want to get into these types of analyses and how we'd interpret the results before we do anything. Regarding allelic mapping bias, it should be enough if we just describe Tuuli's existing analysis better in the paper. - One of the reviewers suggested that dropping some of the analyses might improve the paper. I'm not keen to do that, but we'll discuss this with Magdalena and if she wants us to do that, that's what we'll do. - Gabrielle created a visitor read-only access to the wiki so that the reviewers can go and take a look. So, now it's time to cross your fingers again and hope that our letter convinces Magdalena to reconsider the rejection. We'll let you know when have her decision. best, Tuuli Tuuli Lappalainen, PhD Department of Genetics, Stanford University School of Medicine , and Department of Genetic Medicine and Development, University of Geneva Email:tuuli.e.lappalainen@gmail.com /tlappala(a)stanford.edu Tel: +1 415 351 9713 Bustamante lab Department of Genetics 300 Pasteur Dr. Lane L301 Stanford, CA 94305-5120 USA On /24/113 3:28 AM, Gabrielle Anne Bertier wrote: