You just need to loop it. See the following link:

https://www.google.com/search?client=firefox-b-1&ei=f0xfW8TkKMzTjgS0i7TQDg&q=fisher+test+contingency+table+r+in+loop&oq=fisher+test+contingency+table+r+in+loop&gs_l=psy-ab.3...64027.70691.0.71478.8.8.0.0.0.0.148.805.4j4.8.0....0...1c.1.64.psy-ab..0.5.466...0i22i30k1j33i160k1j33i22i29i30k1j33i21k1.0.4Iwd-_fvUkI

Best wishes, David Booth

Edit: Attached is a file with more options and smarter coding, Fisher2. Not to make this too complicated. It creates a function, and the last line runs the function, so you can put in options in the function call, e.g. multiFisher(file="C:/Users/Salvatore/Desktop/table.txt"), and so on. By default, it sends the output also to file, table.csv, in the working directory, but this can be disabled with saveit=FALSE, or the path can be changed with savefile="D:/myfile.csv". The data frame can be saved as an object for further processing, so that you could do Mydata = multFisher(); hist(Mydata$Fisher.p). What else? I got rid of the weird odds ratio from fisher.test and put in what I think is the common odds ratio. When the mutant is more likely to be diseased, I believe phi is positive and OR1 is positive.

(End edit)

Attached are two files that should work together. It works for a space delimited file, but should work for other types of separators.

Be sure to check the first few lines of results and make sure I built the 2 x 2 tables in the correct order.

Addition: The output includes the statistic called "odds ratio" from the fisher.test function. I'm not sure what it is actually; doesn't appear to be what I think of as odds ratio. If interested, you can consult the function documentation or calculate your own odds ratio.

The output includes the phi statistic, which is helpful to examine as an effect size statistic.

Hints: You can change the file path for "table.txt" in the read.table function.

You could include the table in the R script:

Data

Salvatore,

While I was reviewing your code I noticed that doing the same Fisher test twice is not a good idea, neither is including the same library many times in a for loop.

Recently I had a similar request and here is some part of the code that could help

for(i in 2:11)

{

....

tes

Thanks, Mehmet.

I had the same thoughts when I was writing it. Also makes no sense to re-create the matrix three times. I figured for a few hundred lines and small counts, I would make it as easy to follow (to my mind) at the expense of efficiency.

However, I much prefer the output to be a data frame than directly printed. Perhaps a file output is advantageous, tho.

Maybe I'll clean it up and repost it.

Dear all, Thank you very much for the great suggestions!

David Eugene, I have read the links you sent and I´ve been playing with the " lapply()" function to try to generate loops.

Salvatore S. Mangiafico , I really appreciate your effort on generating the scripts! I was following the codes but I got a couple of errors, I guess I just have to study more to fully understand it.

Jos Feys, thanks for your input, I just couldn´t use all the data I have as a table by writing it, R didn’t recognized it as the data to analyze. Maybe it was too big or a matter of formatting of the file of origin. I am gonna keep on trying.

Mehmet Sinan, thank you for your input also!

Priscila, I added another function with some more functionality, that might help, depending on your trouble. (Probably no different, but at least the new one has a normal odds ratio).

Otherwise you can try to post a piece of your file in the format you'll be using, instead of as a .bmp. It also helps if you start with a small file of a few lines; sometimes there's an irregularity in the file later on that causes the problem.

Please find the link given below , you can get answer :

https://stat.ethz.ch/R-manual/R-devel/library/stats/html/fisher.test.html

For loops are slow and hard to read! Check out the purrr package and the broom package. Something like this:

library(purrr)

library(broom)

data.list %

.$p.value)

Where map_df maps the function (the part after the ~) to every element of data.list and returns a dataframe, tidy() converts the output of fisher.test to a dataframe. Inside of map_df() a "." stands for the current element of data.list, so fisher.test(.) is kindof like fisher.test(data.list[[i]]) if you were doing a for-loop. Similarly, .$p.value just gets the p.value column from the dataframes created by tidy().

Make sure you're doing some sort of p-value adjustment for multiple tests! You can add that in easily with dplyr::mutate() like:

output %>% mutate(adj.p = p.adjust(p.value, method = "fdr"))

Thank you very much, Eric, I will try with these codes!

Hi Priscila Marcondes, I realized that since fisher.test wants 2x2 tables, this is actually a little funky to do. Take a look at this R script. It starts with data that looks like the screenshot you posted and I think does what you want it to do.