GEO datasets

[lsand039]

I ran the MB genes through Gene Ontology using the biological process, molecular function, and cell component annotations. I'm not quite sure how to interpret these results.

[lsand039]

Using Efrain's code, I've been able to find out how many genes are available in the original GSE files. The file below shows the number of genes and probe IDs in the original GSE file, the number of unmatched probe IDs, the number of unique genes, and the proportion of genes included in our analysis.

Gene Counts.png (40.2 KiB) Viewed 39491 times

Gene counts.xlsx

(5.72 KiB) Downloaded 153 times

[lsand039]

These are files in my dataset that used Efrain's code to match the probe IDs and gene names. They match all the genes available in the GPL file to probe IDs in the GSE file much more efficiently than using Access or Base which would usually crash or run out of room.

So far the files I've posted using Efrain's code only contain the raw values of gene expressions. The second tab contains the consolidated/ averaged values of repeated genes where I had to use LibreOffice Calc. I'll be checking and comparing these values to the values I got using Access.

Note: The clinical/ demographic data for GSE15222 and GSE48350 didn't align well in the original series matrix file, so I just grabbed this data from the files cleaned up via Access. I checked the GSM IDs to make sure they matched correctly.

[lsand039]

continuation of the last post

[lsand039]

Here are the files that only the contain the genes common to all the datasets I've used previously. Using Efrain's matching code, I found 8316 common genes, 59 genes more than the 8257 common genes I found using Access. I think Access might not recognized common genes that had a space after them in some GSE files.

[lsand039]

GSE44772 is made up of the same samples in GSE44768, GSE44770, and GSE44771. I'm changing to this file since those three GSE files all come from the same study.

Also attached is the list of 59 common genes that Efrain's code picked up that Access didn't .

[lsand039]

The following contain the Z-score calculations of the all the genes available in the original GSE file.
GSE84422 only contain the definite AD and normal samples.

[lsand039]

A continuation of the last post.

[lsand039]

GSE84422 GPL97 has repeats of GSE84422 GPL96 for the same subjects and the same brain regions, so I've excluded GSE84422 GPL97. There are 7097 genes common in the remaining datasets below

[lsand039]

Continued from last post