Hi there,
I went to download my normalized dataset and I noticed the metadata associations are wrong. I uploaded using time series + one factor, and I made sure the order of samples in the data file was the same as in the metadata files. When I download the data_original file, the metadata is correct. But when I download the data_normalized file the metadata is wrongly associated. (See picture, with original on top, normalized on bottom - I realize the samples are not in the same order but more concerned that the metadata doesn’t match e.g. H05 = concentration H, time 0).
My normalized steps: normalized by median, cube root transformation, Pareto scaling.
Does this wrong matching of my metadata in the normalized file affect the results? Or is it just mislabelled for this download? When I upload as time series + one factor I am not able to edit groups.
Thank you for the note. This is related to a bug in merging data table and metadata table used for preparing final download. The server is scheduled to update during the weekend with the issue fixed, and other updates
It should not affect your results in data analysis - these two tables are always synchronized based on sample names before each analysis.
Thanks for the post and sharing of your data. I noticed the metadata file is pretty large (>5MB) but most “field” of the data file is empty. Some rows should be removed (see screenshot below).
Hi @Qiang, wow I’m not sure how that happened! I have created a shorter version (RBC_metadata_short.csv) and uploaded it to the dropbox folder since I cannot link it here. Let me know if you need it some other way.
Thanks for looking into this!
Sarah