PCA plot no distinct cluster. Only p < 0.05 indicates significant DEGs, while adj p < 0.05 does not.
1
1
Entering edit mode
4 weeks ago
May Ling ▴ 10

Good day to everyone,

I would like to seek advices for my RNA sequencing result. I have 2 groups of samples (10 people for each group) where there are no distinct clusters for PCA plot. According to the DEGs, Only p < 0.05 indicates significant DEGs, whereas no significant DEGs can be detected using adj p < 0.05. Based on my understanding, correcting for multiple comparisons is very important and the adjusted p-value provides a more stringent criterion for significance. May I know if I can proceed with intepreting the DEGs with p<0.05?

RNA-seq padj p-value • 611 views
ADD COMMENT
0
Entering edit mode

Seo et al. (2016), using p<0.05 to determine DEG


Liu et al. (2015), usnig p<0.05 and fold change to determine DEG


ADD REPLY
6
Entering edit mode
4 weeks ago

Never use uncorrected p-values with RNASeq.

Your PCA is suggesting that your sample groups are not very different, and the DEG tests confirm that.

You need to accept what an honest analysis of the data tells you, even if it's not what you like to hear.

ADD COMMENT
0
Entering edit mode

Dear swbarnes2, Thank you for sharing your opinion. I truly value your perspective and insights. However, I have come across some papers, such as Seo et al. (2016) and Liu et al. (2015), where they use a p-value < 0.05 along with a specific fold change to determine differentially expressed genes (DEGs). Could I kindly seek your opinion on why they use a p-value < 0.05 in these instances?

ADD REPLY
1
Entering edit mode

These studies are doing it wrong. If they were basing conclusions of their studies on these DE genes, without further work, they should not have gotten past peer review.

ADD REPLY
1
Entering edit mode

They were presumably doing that because if they did a proper analysis with adjusted p-values, they'd get nothing, and they really wanted to have something.

But what they wanted is not an acceptable reason to deviate from correct data analysis. When you are testing thousands of genes at a time, you must correct for multiple testing.

ADD REPLY
0
Entering edit mode

Dear i.sudbery and swbarnes2, Thank you for taking the time to reply to my message and for providing valuable insights. I now have a clear understanding.

ADD REPLY

Login before adding your answer.

Traffic: 1757 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