FAANG 2022 Workshop Zoom Chats

Session II

04:52:04Ying Wang:@Brenda Great work on sheep. Sorry, probably I didn’t hear clear. How many tissue currently have all 4 histone mark chip-seq datasets for sheep?
04:52:59michael sussman:ISO 20691 covers the formats for FAIR in the life sciences. ISO 23092 covers MPEG-G encoding for the sequences and the metadata.
04:55:10Dominique ROCHA:@Jessica: is there still some "orphan" tissues (not adopted yet)?
04:56:01Jessica Petersen:@Dominique - yes! We have some orphans waiting for adoption. If you are interested, I would be happy to let you know what are available.
04:56:30Brenda Murdoch:@Ying we have 47 tissues with all four histone marks.
04:57:17Jessica Petersen:Might not have made that last comment available to all:
04:59:19Ted Kalbfleisch:@Dailu. We’ve loaded about 8 datasets before, but yes, if you are looking at a region with lots of expression, it will slow down.
05:20:02Doreen Becker:@Wesley: How many marker genes for inferring cell type do you usually use?
05:20:25Dailu Guan:Do you consider doing deconvolution using your scRNAseq data?
05:20:38Michèle Tixier-Boichard:Many thanks Wes. One practical question: are your protocols for nuclei isolation from specific tissues available on the FAANG data portal ?
05:21:02Dominique ROCHA:@Wes: as you start with nuclei, will you do scRNA/scATAC seq for the same cell preps as well, to connect genes expressed and regulatory regions?
05:26:26Wesley Warren:We typically use the top 20 DEGs per cluster to identify cell type. Once we have robust cell type markers we plan to test some deconvolution algorithms with our cell type specific training set. We can place our nuclei prep protocols in FAANG but will also publish these as part of an initial immune cell atlas. We don’t plan to generate scATACseq data at this point on these samples.
05:30:09Dominique ROCHA:@Lingzhao: how did you compare genetic variants/RNA-seq expression diversity (across breeds)? Did you use only one tissue (i.e. blood or muscle) or did you combine several?
05:36:39Dominique ROCHA:@Lingzhao: will you make public the RNA-seq data (count) like you did for cattle with the Cattle Gene Atlas?
05:39:50Andreas Pfenning:@Lingzhao, Amazing resource! Did you compare orthologous genes or loci to the human GTEX project? Do observe/expect some similarity across that evolutionary distance?
05:41:55Andreas Pfenning:@Lingzhao, very cool, thanks!
05:44:15Dailu Guan:@Lingzhao, Great talk!!
05:47:55FANG Lingzhao:@Dominique: Thank you very much for the question. when comparing eQTLs/gene expression, we only focused on one tissue at a time. Yes, all the data from FarmGTEx will be publically available later
05:48:24Dominique ROCHA:@Lingzhao: super. :)
05:52:14FANG Lingzhao:@ Andreas, Yes, we compared the orthologous genes and loci between humans and pigs. We did observe certain conservation of eGenes and eVariants between these two species. In the future, we will systemically compare gene expression and eQTLs across all the farm animals analysed in FarmGTEx and humans.
05:52:50FANG Lingzhao:@Dailu, Thank you very much
05:54:44Dailu Guan:@RuiDong, I am wondering what the conserved SNPs mean in your context? If it means a SNP in cattle is still a SNP in human, for instance?
05:56:31David Hawkins:Unfortunately, I’ll miss the round table discussion. I have to teach. Great talks everyone!
06:02:49Dominique ROCHA:@Ruidong: will the 50K functionaly enriched variants will work for beef cattle too?
06:03:58Robert Mukiibi:@Ruidong How would such a functional panel perform for traits that do not have major QTLs eg Feed efficiency?
06:03:59Dominique ROCHA:@Ruidong: overlap between QTL/eQTL is up to 10% but what about splicing QTLs. Looks like that mQTLs and sQTLs were better than eQTLs in your initial work (PNAS)?
06:04:12Christopher Tuggle:@Ruidong- how does this use of FANAG type data compare with your earlier work showing evolutionarily conserved positions add to the prediction accuracy?
06:05:01Mazdak Salavati:@Ruidong What are your thoughts on milk focused costume chips and selecting high producing animals at the expense of loss of longevity or welfare estate of the dairy cattle?
06:10:25Christopher Tuggle:AG2PI could be interested in bringing these groups together to help USDA understand how to organize future RFAs in G2P
06:10:33RuidongXiang:@Dominique: yes we do see 'better' results from sQTL, this is because sQTL mapping is more powerful, so they deliver more informative variants to be used in genomic mapping/selection; mQTL is also good, but at the moment we don't have very large sample size to increase the power of detecting them, we are still working on evaluating the merit of them.
06:12:07Dominique ROCHA:@Ruidong: thanks for your reply. Is your list of finally mapped pleiotropic variants available?
06:13:15Christa Kühn:The challenge for the global FAANG will be to harmonize across the major funding agencies USDA, EU etc., particularly for larger Projects beyond travel etc.
06:13:56RuidongXiang:@Christopher: there are a lot of types and large volume of new FAANG type data coming and we are in the process of assessing these datasets again. The conserved regions are still very competitively enriched in heritability, but we want to see if new FAANG datasets can change this picture.
06:16:59James Koltes:@Emily Clark: If folks are interested in the AG2PI data reuse meeting that Chris and I lead, a summary of that meeting is at our website: https://www.ag2pi.org/workshops-and-activities/community-workshop-2022-02-09/ There are recordings available from our speakers and we will add more information overtime. Unfortunately, I need to run, but I wanted to share this for those who may be interested.
06:17:02RuidongXiang:@Mazdak: that's a very good question. In fact the customised chip is based on 35 traits, so they included non-milk traits like fertility and health. However the benefit of the customised chip has biases towards milk traits because they have much larger sample size and this is true for many organisations. so to improve this will probably need international collaborations to increase the sample size of these traits with fewer records
06:18:59Christopher Tuggle:@Ruidong- thanks! It would be interesting to see if functional annotation WITHIN species can overcome the evolutionary conservation, which is so effective.
06:19:06RuidongXiang:@Dominique: yes! here are they: https://figshare.com/s/93bd992a42786f9466b7 . the coordinates are based on UMD3.1 but we have tested liftovering to ARS and it works.
06:19:21Mazdak Salavati:@Ruidong Thanks for your answer. Indeed focus on the dairy journey of the animal is a very difficult task but surely going forward we should try to bring more longevity phenotype recording on board chip designs. Otherwise we will end up repeating the history for the black and white cow. Very nice talk and thanks for elaborating.
06:19:26Robert Mukiibi:Will there be special sessions dedicated to FAANG projects at the WCGALP?
06:19:59Dominique ROCHA:@Ruidong: thanks
06:20:35Robert Mukiibi:Thank you.
06:20:35Christa Kühn:There is a compendium of talks in 3 subsessions, taking About 4 Hours at the WCGALP
06:21:23Michèle Tixier-Boichard:Regarding bioinformatics structure, it seems that there are several parallel initiatives, when do they merge (GTex for example, what Ted presented?)
06:23:06RuidongXiang:@Christopher: we hypothesise that the across-species conserved and within-species annotation may be both important but function a bit differently, it would be good to line them up to do direct comparisons.
06:25:36RuidongXiang:@Mazdak: yes agreed. I think some of the presentations talking about DNA methylation related to longevity are really cool. They may be other angles from functional genomics to improve non-production activities.
06:25:49Meenu Bhati:@Ruidong, have you compared mammalian conserved score vs conserved elements in 100 vertebrate ? Considering more similar function evolution than whole vertebrate.
06:27:58Fiona McCarthy:I think it is important to have specific pipelines and that does not stop people from having/running their own pipelines. Standard pipelines/workflows will allow benchmarking.
06:29:53Christopher Tuggle:I agree with Fiona (and Ted). Who wants to volunteer to organize a discussion
06:30:16Ted Kalbfleisch:I’m happy to volunteer.
06:30:31Fiona McCarthy:Me too
06:30:49Christa Kühn:nf-core could be a great starting Point with the DSL2 modulation options
06:31:40Fiona McCarthy:@KristaKuhn Agreed! Would like to see discussion before additional/new workflows are added.
06:32:11Fiona McCarthy:Have the subgroups (e.g., bioinformatics) been meeting?
06:35:59Christopher Tuggle:I think Mick as Chair of Bioinformatics has not been calling meetings, but I don’t know who he invites to such meetings.
06:37:26Ole Madsen:As far as know there have not been meetings for years
06:37:56Peter Harrison:@Fiona. The pandemic did really slow many of the global subgroups. As this survey highlights, time to relaunch and refresh the scope of existing subgroups and launch new ones.
06:38:15RuidongXiang:@Neenu: yes we have compared conserved across 32 mammals VS 100 vertebrates in terms of heritability enrichment. conserved 100way is slightly more enriched than the conserved 30way, but they are both very strongly enriched. Our previous work (https://www.pnas.org/content/116/39/19398.short, in the supplementary data) has showed this.
06:38:52Mazdak Salavati:A bioinformatics focused subgroup meeting is definitely needed. Lots has happend that can benefit from practitioners feedback.
06:41:10Amanda Chamberlain:There is a lot of data now so maybe a large combined data or meta analysis would be a good way to get groups working together and producing high impact papers
06:42:16Fiona McCarthy:Not than I am aware
06:42:16Amanda Chamberlain:No subcommittee meetigns for ages
06:43:16Christa Kühn:Maybe, we refresh the subcommittees with input from bioinformaticians AND data producers
06:43:23Amanda Chamberlain:There wasn't any data to work on so it was speculative pipeline talk
06:43:38Peter Harrison:A lot of the meetings were before their time, we now have a lot of highly active data and people funded to work on pipelines and other aspects that can therefore contribute more.
06:43:45Fiona McCarthy:@AmandaChamberlain Agreed!
06:44:58Emily Clark:When I was an early career researcher the sub-committees, particularly the metadata one, were the way I became really involved in FAANG 🙂
06:47:41Amanda Chamberlain:Will the posters and talks be available after?
06:48:33Mazdak Salavati:👏🏼 Thanks for organising the workshop.
06:49:30FANG Lingzhao:Thanks a lot for organizing this exciting workshop!
06:49:36Dominique ROCHA:Thanks. Interesting talks and posters. Many thanks to all presenting. How a good day/night/lunch/dinner.
06:49:38Ryan Corbett:Thanks everyone! Great workshop
06:49:47Androniki Psifidi:Thanks everyone :-)
06:49:56RuidongXiang:excellent session! thanks everyone!
06:49:59Oladipupo Bello:Thanks
06:50:13Smahane CHALABI:thanks!
// end //