3

I have a proteomics dataset in which I compare two conditions, and therefore I have condition A (8 replicates) and condition B(9 replicates).

In proteomics we don't always have measures for each replicate, but we don't want to eliminate a protein observation (each row is a protein) because 1 of 8 replicas is presenting a NA value.

I would like to use a t test and I have tried the function na.action = "na.pass" but the following error arises: 

 Error in t.test.default(x[1:8], x[9:17], paired = FALSE, na.action = "na.pass") : 
  not enough 'x' observations

I give you an example of the data set:

        C1_1  C1_2  C1_3  C1_4  C1_5  C1_6  C1_7  C1_8  C2_1 C2_2   C2_3  C2_4  C2_5  C2_6  C2_7  C2_8   C2_9
Prot_1  4.42    NA  0.73  0.47  0.84    NA  1.24  0.99 1.21  0.84   0.64  0.61  0.76  0.29  0.88  4.09   1.15
Prot_2  0.56  0.30  0.30  0.27  0.18  0.48  0.59  0.20 0.36  0.40   0.44  0.49  0.33  0.67  0.30  0.50   0.75

I would be very grateful for your help,

Best Wishes,

Julia

Roland
  • 127,288
  • 10
  • 191
  • 288
Julia Bauza
  • 31
  • 2
  • @R-Fever thanks for the info, looks like I was wrong. – ytk Dec 15 '16 at 16:46
  • I'm unable to replicate the issue with the data you provided, the code is working – ytk Dec 15 '16 at 16:50
  • 4
    This may be more of a statistical issue masquerading as an R coding issue. What do you want to know about your data? What are the `Cj_i`s? Are the data essentially paired? What would it mean for NAs to be passed? – gung - Reinstate Monica Dec 15 '16 at 16:51

0 Answers0