How to find up-regulated and down-regulated genes after GEO2R analysis?

Question

[DanielC]

Dear Friends,

I performed GEO2R analysis on a series dataset "GSE57475", and got the below sample values from the GEO2R analysis (from the GEO website). Can you please suggest on how to find the up-regulated and down-regulated genes in a series file like "GSE57475". Thanks much!

ID  adj.P.Val   P.Value t   B   logFC   Gene.symbol Gene.title
ILMN_1734855    0.636   0.000013    -4.5    2.228   -0.3797 SELPLG  selectin P ligand
ILMN_1675124    1   0.0000648   -4.11   1.016   -0.5205 DDX17   DEAD-box helicase 17
ILMN_2150654    1   0.0002003   -3.81   0.165   -0.1422 ZSWIM4  zinc finger SWIM-type containing 4
ILMN_2204754    1   0.0004128   -3.61   -0.377  -0.2404 TMX4    thioredoxin related transmembrane protein 4
ILMN_1832425    1   0.0005167   -3.55   -0.545  -0.164      
ILMN_2388955    1   0.0008171   -3.41   -0.887  -0.3191 FAM212B family with sequence similarity 212 member B

Regards, DK

[arta]

To find up-regulated and down-regulated genes, you should check logFC (Fold-change in log2 scale (generally)). consider value of 1 in log2 is 0. There is optimal cutoff but log2 > 1 indicates up-regulation and log2 < -1 indicates down-regulation of genes. Moreover, you should consider adj.p.val which is adjusted p value (corrected p value dues multiple comparison). Again there is no generally accepted cutoff but you should consider values < 0.05 which indicates the test is statistically significant.

For example the data you pasted, there is no significant genes since there is no genes adj.p.val < 0.05 and |logFC| > 1. Hope the rests do not look like those.

[DanielC]

Thanks much! Actually in this particular GSE series none of the adj.p.value is < 0.05 - they are all "1" and a "0.636". However, the logFC values are >1, but none of the samples have a condition of "p <0.05 & logFC > 1". So, can it be said that in this case, there are no up-regulated and down-regulated genes or differentially expressed genes (DEGs)? Is there any other way to find out the DEGs, up-regulated and down-regulated genes in a series GEO file? Should I check whether the series is "normalized" or not, may be that is the reason?

Thanks much for your time. DK

[arta]

How many genes are in the data ? Are you also sure that you compare the different conditions such as WT vs KO or Tumor vs Normal ? Maybe you should check how the data were preprocessed.

[DanielC]

Yes, I selected the tumor and the normal for the geo2r analysis - there are two conditions in this dataset "normal and tumor" which I selected and performed geo2r. There are 19000 genes in the dataset. Do you think this could be a good dataset to do computational study based on the preliminary results? Thanks much!

[h.mon]

Is there any other way to find out the DEGs, up-regulated and down-regulated genes in a series GEO file? Should I check whether the series is "normalized" or not, may be that is the reason?

We don't know what you did exactly, so how could we know?

I performed GEO2R analysis on a series dataset "GSE57475"

This is very vague, post your code if you want more informed help.

[DanielC]

Thanks for the reply. I performed the analysis using the GEO website (using "Analyze using GEO2R"), so there is no code.

[zizigolu]

up <- which(results$logFC == 1)