Looking for replicated conditions

[IsBeth]

Hello! I'm a newbie and I'm trying to write a function which can read how much replicated conditions there are in an experimental design in a given case. For example, if I have a design (for example an input like a data.frame) like the following:

sample condition
1 C1.sam A
2 C2.sam A
3 C2.sam A
4 Infect1.sam B
5 Infect2.sam B
6 Infect3.sam B

How can I access the number of replicates? (perhaps some function similar to apply?) Cause obviously there are three replicates called "A" and three replicates called "B", but I don´t know how to store this in a variable, so I can work with that information later on the script.

Thanks and sorry for my beginner's English

[dpryan]

It sort of depends on how you want to use things later on. If the dataframe is called "d", then:

table(d)

That will give you a table with the conditions as names and counts as values. In general, though, you can just "~d$condition" or use model.matrix.

[IsBeth]

Thank u dpryan. The table-function worked =)

What if I wanted to use a function like that with my counts, not with my design table? A data frame like this for example:

Control1 Control2 Control3 Infect1 Infect2 Infect3
4 34 23 12 2 33
6 14 15 17 32 22

I want to get the information that the condition "control" is replicated 3 times, and condition "infect" is replicated 3 times, too (so that R get just the information in common ignoring the numbers).

[dpryan]

There's no simple function for that since R would have to already know your general labeling format. Presuming that numbers at the end of sample names can always be stripped, then the general idea would be to:

1) Save the names in another variable (foo <- names(d))
2) Remove numbers from the end (normally this would be a regex)
3) You could then use things like "table" or whatever on the results.

Of course, if you new ahead of time what the groups were you could just count how many contain a phrase (i.e., use grep), but I'm assuming you want to do this as part of a general purpose pipeline.