I have a dataset that has abundance of different bacteria in different mice type. Let’s say mice are called: sw/c57, sw/sw, c57/c57 and c57/sw. The fecal samples of these mice have been analyzed and the number of present bacteria in their fecal samples is available. Now I want to find those bacteria that are significantly different in any of the mice type. E.g. Prevotela is significantly different in c57/c57 comparing to other mice.

An example of my dataset:

           sw/sw    sw/c57    c57/c57    c57/sw
Roseburia  500      141       0          589
Prevotela  300      131       237        288

Each mice type has 10 replicates.

What kind of statistical test I should use to analyze my data?


  • $\begingroup$ Are you trying to compare proportions? $\endgroup$ – Glen_b Jun 27 '14 at 10:31
  • $\begingroup$ No I am comparing the absolute values. But I don't know if I should compare proportions instead? $\endgroup$ – Rozita Akrami Jun 27 '14 at 11:47
  • $\begingroup$ The main question is if we can find any differences between mice in any bacteria. $\endgroup$ – Rozita Akrami Jun 27 '14 at 11:47
  • $\begingroup$ Whether you are interested in the raw counts or proportions is up to you - it depends on the specific question you're trying to answer. It sounds like a difference even in the total count across mice would matter. $\endgroup$ – Glen_b Jun 27 '14 at 16:00

Perhaps I am missing something but if you are interested in the total count for each type of bacteria, why couldn't you do some form of count regression? Probably negative binomial regression with count as the DV and mouse type as the IV. You could then compare each pair of mice.

Whether to correct for multiple comparisons will be an interesting question.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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