0
$\begingroup$

I have data that looks like this:

  selected_genes <- c("A","X","Y")
  pooled_genes <- c("A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", 
"M", "N", "O", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", 
"Z")

The question I want to answer is that if we pick 3 genes from pooled_genes how likely that selected_genes is obtained.

What's the appropriate statistical test to obtain p-value for that?

Improve this question
$\endgroup$
7
  • 2
    $\begingroup$ Does the order of selected_genes play a role? Can genes be picked twice? Can selected_genes contain the same gene twice? Please extend your question, so that answers are either right or wrong and not depending on interpretation. $\endgroup$
    – Bernhard
    Commented Mar 14, 2018 at 7:29
  • 1
    $\begingroup$ @Bernhard Thanks a lot. Order of selected genes doesn't matter. Yes the genes can be picked twice. No selected genes cannot contain the same gene twice. $\endgroup$
    – neversaint
    Commented Mar 14, 2018 at 7:31
  • 1
    $\begingroup$ Oh sorry for the second comment. As "A" is in the selected_genes but not in the pooled_genes, the answer is easy: It is not likely at all. $\endgroup$
    – Bernhard
    Commented Mar 14, 2018 at 7:31
  • 1
    $\begingroup$ @Bernhard Can you elaborate. We (my boss) need a p-value to justify. $\endgroup$
    – neversaint
    Commented Mar 14, 2018 at 7:32
  • 1
    $\begingroup$ @neversaint, what Bernhard meant is that gene "A" in "selected_genes" does not occur in "pooled_genes", thus "A" cannot be picked, and the corresponding probability is zero. $\endgroup$
    – crlb
    Commented Mar 14, 2018 at 7:51

1 Answer 1

Reset to default
1
$\begingroup$

Assuming the order of picking does not matter, and genes are sampled with replacement from pooled genes (you dont delete a sample from pooled genes after it has been picked), the corresponding distribution is the Multinomial distribution.

From the distribution and assigning probabilities of picking genes, $p_i$, $i \in \{A,B,...\}$, the p-value can be calculated given an observation.

Improve this answer
$\endgroup$
1
  • $\begingroup$ c("A","X","Y") is the observation. How can I get the p-value with R code? $\endgroup$
    – neversaint
    Commented Mar 14, 2018 at 8:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.