How to treat data with three biological replicates

Question

[mengmpp]

Hi all,

for each point (control and treatment), I have two more biological replicates, like ,control: A0,A3,A6, treatment:A1,A4,A7. After normalized with RPKM, I got a set of data with log2(A10/A0), P-value, FDR; log2(A4/A3), P-value, FDR; and log2(A7/A6), P-value, FDR. After that, I found common genes which consistently expressed in three replicates. That is main aim I want to extract. OK, the problem is for these common gene, they have three set of data, how can I assemble to one set?

Thanks a million.

[mengmpp]

A0-Expression     A1-Expression     A0-RPKM      A1-RPKM      log2 Ratio(A1/A0)     P-value       FDR           A3-Expression     A4-Expression     A3-RPKM      A4-RPKM      log2 Ratio(A4/A3)     P-value       FDR           A6-Expression     A7-Expression     A6-RPKM      A7-RPKM     log2 Ratio(A7/A6)     P-value      FDR
715               79                45.79892     4.967396     -3.20475              7.76E-132     1.34E-129     938               113               60.01701     7.530592     -2.99454              2.05E-155     2.88E-153     365               30                23.13004     2.03935     -3.50359              1.47E-69     1.35E-67

[GouthamAtla]

Which method/tool did you use to get the P-value or FC? I think you have treated them as individual samples rather than replicates.

[mengmpp]

Yes, I got these data from company, I think they treated them as individual samples.

For Quantification of gene expression, use RPKM algorithm; for Screening of differentially expressed genes, use Poisson distribution model; for Screening of group differentially expressed genes, use NOIseq method

[GouthamAtla]

Just try it urself with DESeq/edgeR. They are well documented.

[Learner]

@mengmpp did you find a solution for this ? I am facing the same problem

Answer 1

[Ido Tamir]

How do you get a p-value and an FDR from the comparison of 2 data points? llike: 0.4 RPKM vs 4 RPKM?

Please use a tool like DESeq2 or edgeR. They do everything automatically.

http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html