search

diwis / BUFET

Boosting the Unbiased Functional Enrichment Analysis
GNU General Public License v3.0
6 stars 1 forks source link

Calculation of Observed.Target.Gene.Overlap.Proportio #2

Open ralodo93 opened 3 years ago

ralodo93 commented 3 years ago

We were trying to understand how do you calculate the Observed.Target.Gene.Overlap.Proportio by doing the operations by hand. Let me provide a concrete term example:

This is our results:

GO.term.ID GO.term.size Observed.Target.Gene.Overlap.Proportio Mean.Random.Simulated.MicroRNA.Target.Overlap.Proportion
hsa03010~Ribosome           27                             0.00276243                                               0.00445511
   One.sided.empirical.p.value Benjamini.Hochberg.0.05.FDR
                   0.345034                           -

This is involved-genes.txt

>hsa03010
MIR1    
MIR2    GEN1,GEN2
MIR3    
MIR4
MIR5    
MIR6    
MIR7    
MIR8    
MIR9    
MIR10

What we are expecting is Observed.Target.Gene.Overlap.Proportio = 2 / 27 Being 2 (GEN1,GEN2) the number of target genes and 27 the number of genes annotated in hsa03010. But 2/27 (0.07407407) is not the reported Observed.Target.Gene.Overlap.Proportio (0.00276243).

zagganas commented 3 years ago

As specified by Bleazard et al., the overlap proportion is the number of the group's targets targets participating in the process divided by the total number of the targets. This means that in your case, it would be 2/n where n is the union of genes targeted by MIR1-MIR10.

zagganas commented 3 years ago

Also n=2/0.00276243 which is about 724 genes.