P value for test for proportions and random sampling
0
0
Entering edit mode
2.7 years ago
pixie@bioinfo ★ 1.5k

Hello, I have a gene list with Up=208, Down=135 and No change=44. Total =387

I have done a random shuffling 1000 times and obtained the mean Up=77.54, Down=74.268 and No change= 235.19

I want to test if the no. of Down-regulated genes are significantly different or Not from the random sampling. Is a test for proportions applicable here ? Any other suggestion ? I would prefer to have stricter conditions because our hypothesis was to show that the proportion is close to the random sampling.

In R:
prop.test(135, 387, 0.19, alternative = "two.sided",correct = TRUE, conf.level = 0.99)


Kindly suggest. Thanks

R • 449 views
0
Entering edit mode

Can you explain a little more? These genes were not selected from a larger universe? Are you shuffling only the 387 genes? Or are you shuffling a larger universe and then selecting 387? i.e. you shuffle your universe 1000 times, and then draw a pool of 387 which contains a mean of 77 up, 74 down, and 235 no change? (usually gene sets are drawn from a larger universe. can you describe your numbers a little more?)