chi square test providing expected data in R

Question

First time writing in here, so I apologize if something is missing.

I am comparing two DNA quantification methods and I am trying to see which method gives me results closer to the real ones (the DNA that I added to the samples).

For each method I have 5 replicates. I put 1000 cells in each sample (expected value) and got the following quantification values (observed values):

Method P - 500, 400, 400, 500, 500
Method Q - 1000, 900, 1400, 700, 1000

When I try to use the chisq() function, I appear not to be able to tell the function which are the expected values, it calculates the expected values and that is not what I want.

> P<-c(500, 400, 400, 500, 500)

> Q<-c(1000, 900, 1400, 700, 1000)

> chisqQ <- chisq.test(Q)
Chi-squared test for given probabilities

data:  Q
X-squared = 260, df = 4, p-value < 2.2e-16

> chisqP <- chisq.test(P)

    Chi-squared test for given probabilities

data:  P
X-squared = 26.087, df = 4, p-value = 3.039e-05

The problem with this is that I do not establish my expected values, and while for the Q it automatically calculates 1000, for the P it doesn't

> round(chisqQ$expected,2)
[1] 1000 1000 1000 1000 1000

> round(chisqP$expected,2)
[1] 460 460 460 460 460

There is the p argument in the chisq function, but it has to be a probability, which is not my case.

I have calculated the chi square values by hand on excel and compare them, but once, in the future, I will have several techniques and several cell amounts I would like to know if it is possible to do in R.

Thanks in advance,

Cheers,

Joana

Give us some reproducible data or example, and show the codes you wrote so far... That will help you to get answers... — Thai

ka0 ka0 · Accepted Answer · 2017-10-20T13:09:07

So if I understand this correctly, you already have the expected values and want to use chi square to see how good of a fit you have.

If so the following solution will work.

obs <- c(500,400,400,500,500)
exp <- c(XX, XX, XX, XX, XX)
chisq.test(x = observed, p = expected)

chi square test providing expected data in R

1 Answers