Performance is very similar across samples when using their respective LOO statistics, and the new Tomsk dataset performs equally well everywhere. In contrast, naive PGS testing for Moscow and Ufa with the original PGC3 sumstat looks horribly optimistic.

To resolve causal variants in meta-analysis, studies typically apply sumstat-based fine-mapping as if they were a single cohort. However, it's unclear whether heterogeneity across cohorts (e.g. sample size, phenotyping or imputation) affect fine-mapping calibration & recall 2/n

To mitigate this, we developed SLALOM (Suspicious Loci AnaLysis Of Meta-analysis summary statistics), a sumstat-based QC method that identifies suspicious loci for meta-analysis fine-mapping by detecting association outliers based on local LD structure 5/n https://t.co/4rzXpQs363

Why do you even want to store GWAS sumstat in a VCF format?

@pastramimachine @kdc509 Buzzy new sumstat

Ketika temanku sibuk sumstat freq/cumfreq aku sibuk extract garch variance

Check our NEW #SumStat format to increase usability! Our manuscript https://t.co/uBOrx1i1DZ GitHub repository https://t.co/8uimQYbCKd All details on our blog post https://t.co/35GyZKi0vA Share your data https://t.co/1wHeyIeASk #OpenAccess #FAIRprinciples 📷

@SbotGwa @astheeggeggs @bmneale @uk_biobank Highlights include expanding to over 4,000 phenotypes, new assessments of confidence and statistical significance, empirically estimated power curves for LDSC, and LDSC-formatted sumstat file downloads. Browse it all at https://t.co/xNumiX56uq

P2S 29 – Converting files from Plink to Sumstat format http://t.co/tqxqDMM2

P2S 20150226 – Converting files from Plink to Sumstat format http://t.co/Pd4Oo7pkWG

@privefl Over the 200 random phenotype vectors, the expected correlation between the sumstat i and the sumstat j is the (i, j)-element of the LD matrix. If you have some bad quality prior guess about the matrix, maybe you can do a Bayesian update on that.

SUMSTAT / SumstatQ – Sum Statistics for multi-locus Genetic Association Mapping http://t.co/7HatYnbs

Sumstat On Nikki Doe

@ErrataRob Also, will add bro scripts to start logging all http headers; perhaps a sumstat to perform some basic freq analysis

@NoCHASERR_J lol That's Chief Keef Not Pops ... Sumstat

@michelnivard @AlexTISYoung Yes, it should be a weighted average (just like each sumstat is a weighted average of site-specific sumstats). The idea is to mimic what would have happened if we analyzed all the individuals in all of the cohorts jointly.

Alvarado's spatial sumstat magic: a spatial PCA step #Evol2014 #phylogeography

@timrpeterson @bmneale @Dropbox @AWS Hi Tim - following up here as we are still in limbo with Dropbox. If you have a google account we can share via a google bucket for the time being. Please email nealelab.ukb@gmail.com with account and sumstat list and we can try to get something set up

🚨 Super excited to see Tsz Fung's preprint out! Together with @CharlestonCWKC, we developed a sumstat-based approach to quantify heritability explained by local ancestry markers. This has been a massive effort by Tsz Fung and I'm so proud to see this out! Details below 👇