Integration of RNA-Seq data with microarray data
1
3
Entering edit mode
4.7 years ago
Fahimeh.6868 ▴ 30

Hi everyone, I'm researching growth Differential Expression Genes (DEGs). To do this, I chose fetal , juvenile and adult ages. I found the fetal age data from the GEO site, which was prepared by microarray method and Juvenile and adult ages data from the ArrayExpress site which was prepared by rnaseq method. I calculated the expression values of the microarray and rnaseq data separately. then I agrigate and merge them together based on the gene symbol, next I removed the batch effects. finally I got the differential expression genes. At each stage, the plot was drawn, which showed that the process was correct. Do you think my work process was right? Isn't it a problem that one age is with the microarray data and the other two ages is with the rnaseq data?

microarray RNA-Seq batch-effect R • 2.3k views
ADD COMMENT
6
Entering edit mode
4.7 years ago
lessismore ★ 1.4k

You have two problems:
1. the dynamic ranges of microarray and RNA-seq are different. I suggest to read this to normalize it https://www.ncbi.nlm.nih.gov/pubmed/26844019
2. the technical confounding. You don't know if your groups cluster correctly because of biological reasons or technical reasons. For correcting for technical confoundings you should have samples from different experiments you want to integrate in each of the age group that you have. I suggest this reading https://www.ncbi.nlm.nih.gov/pubmed/27540268

ADD COMMENT
0
Entering edit mode

Hi , In analyzing with DeSeq2 package, we have a function that converts read counts to continuous data, (below command) cnt2<- counts(cds, normalized= T). This is my question now: can I extract cnt2 and merging it with an expression matrix of microarray? And then I remove the batch effect by sva package. Is this method correct?

ADD REPLY
0
Entering edit mode

No i don't recommend to do this. I suggest to go for a quantile normalization. You can read about it In the paper at the point 1 i suggested. It's not difficult. Briefly, you order microarray and RNA-seq values. Then you replace RNA-seq values with the microarray ones. In this way you're sure that the dynamic ranges are the same and the data is comparable.

ADD REPLY
0
Entering edit mode
ADD REPLY

Login before adding your answer.

Traffic: 1635 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