Is there any statistical test for comparison of two lists of proteins ?

Question

[ma23]

Hi all, I have two lists of proteins: one obtained from a healthy person's sample and the other - from a diseased one. Each protein has a score - a number that represents the degree of confidence that a protein really was represented in a sample. I want to make some kind of comparison between these two list to find if there are any differentially represented proteins. How can I do that ? May be there are some statistical tests that I could use ?

[Kevin Blighe]

a number that represents the degree of confidence that a protein really was represented in a sample

Please explain some more about this? How is this score derived?

[shoujun.gu]

This is similar to DE analysis. But with one sample in each condition, its statistical power would be very poor.

Answer 1

[Asaf]

Short answer: you can't. Not with one sample of each condition. Statistics is taking a subsample and studying about the entire population and with only one sample it's quite not possible.

However, you can always look at proteins that are different in both samples and try to make a claim about this group of proteins, for instance that a lot of them appear in a specific pathway or a lot of them (statistically significantly large number) are regulated by the same transcription factor etc.