p value and adjusted p value in microarray data analysis
5
0
Entering edit mode
6.2 years ago
7794karthi • 0

i am doing microarray data analysis using R package. my doubt is regarding selection of DEG! whether i should consider p VALUE <0.05 or ADJUSTED p VALUE < 0.05 for selection of DEG?? i get adj p values >0.3. what should i do now? on what basis the threshold of p value or adj p value should be set??

 

R • 23k views
ADD COMMENT
0
Entering edit mode

yes,I got all those p values and adjusted p values. But the problem is , all adjusted p values are greater than 0.3. so how should i choose DEG?

ADD REPLY
2
Entering edit mode
6.2 years ago
Sam ★ 3.5k

You should use the adjusted p-value to account for multiple testing. You should also know what kind of adjustment was done, be it FDR or bonferroni. There is no absolute "best" p-value threshold, we usually just use 0.05 or 0.01 as a common practice 

ADD COMMENT
0
Entering edit mode

If you use adjusted p-value, then you should use 0.05 or 0.01. However, if you use the un-adjusted p-value, you should use an adjusted threshold which should be 0.05/(number of tests). This is the bonferroni correction

ADD REPLY
0
Entering edit mode

You can just use the add reply button instead of generating new results.

For any analysis you do, you should be able to get the p-value and the corresponding adjusted p-value. I am not sure what packages you use. But generally, you can just use the p.adjust function in R to get the adjusted p-value. You can use either the FDR or bonferroni correction

ADD REPLY
0
Entering edit mode

yes,I got all those p values and adjusted p values. But the problem is , all adjusted p values are greater than 0.3. so how should i choose DEG?

ADD REPLY
0
Entering edit mode

It means that after corrected for multiple testing, you have no significance. Sometimes, there just isn't any signficant finding /DEGs in the study. 

Or you can try to use FDR. If the multiple testing is based on Bonferroni, it will be more conserved and might therefore lead to very few findings

ADD REPLY
0
Entering edit mode

than you sir..

ADD REPLY
0
Entering edit mode
6.2 years ago
andrew ▴ 520

As Sam mentioned, if your smallest p-value is 0.3, then it means you do not have any significant DE genes.  Did you perform any QC or normalization before calculating differential expression?

In our application, iPathwayGuide, we recommend DEG selection be based on adjusted p-value (and log2FC).  Our default thresholds for DEG selection are Adj.pv < 0.05  and absolute value log2FC > 0.6.

If there are no DEGs, then it means you might need to re-run some of your experiments or that the two phenotypes are not dissimilar.

I hope this helps.

ADD COMMENT
0
Entering edit mode

Thank you sir.. i have understood the concept

 

ADD REPLY
0
Entering edit mode

Sir, what is difference between adjusted p-value and p-value? which one is better to use?

ADD REPLY
0
Entering edit mode
4.2 years ago

Think you.you help me.

ADD COMMENT

Login before adding your answer.

Traffic: 1649 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6