What kind of data? How many samples? And so on. There are many very good packages with lots of documentation around - like DESeq2 and edgeR - and I feel reading them carefully will answer your questions.
If you only have 2 samples, there is a minimal amount of things that you can do. You should still normalise them together and then just look at, for example, ratio of expression for your gene of interest. p-values for a 1 vs 1 comparison make little sense - you would be laughed out of the room if you presented p-values for n=2
How To Ask Good Questions On Technical And Scientific Forums
What kind of data? How many samples? And so on. There are many very good packages with lots of documentation around - like DESeq2 and edgeR - and I feel reading them carefully will answer your questions.