Hi, I just wanted to confirm that what covariates were used in the differential expression results for the AD diagnosis model that's available on AGORA. I know that the RNA-seq harmonization study has MAYO, MSBB, and ROSMAP studies in it. The study details here provide the preprocessing steps and the R package sageseqr that was used for differential expression (https://adknowledgeportal.synapse.org/Explore/Studies/DetailsPage/StudyDetails?Study=syn21241740). The individual covariate files are also available for all three studies ( https://www.synapse.org/#!Synapse:syn29855549). It's also mentioned that sageseqr empirically selects meaningful covariates however it's not mentioned which covariates were eventually selected for the results displayed on AGORA. Can you please help me with that? Thank you! Best, Zainab

Created by zainab khurshid zainabk
Thank you so much! This is very useful.
Hi @zainabk, Sageseqr identifies significant covariates and selects which ones to put in the model, based on whether the model is improved via BIC. You can find metrics for covariate significance and the best model identified in the html files (look for the heading "Model Identification" to find the model that was actually used in the analysis). MSBB file is [here](syn27068766), Mayo file is [here](syn27024974) and ROSMAP file is [here](syn26967461). These files also contain lots of other interesting information, like summary metrics and outlier identification. I hope that helps! Best, Laura

The RNA-SEQ harmonization study AGORA RNA-SEQ results page is loading…