sample gene count matrix from list in R

[tonybert]

Greetings, I am trying to produce a sample matrix consisting of samples (columns) and gene-IDs (rows), with the raw hits for each gene in the matrix. For COG categories, I have been able to produce these types of tables quite easily in excel. However, for other types of gene-ontology databases that are a bit more complex, i have been running into some problems. What i have currently is a list that looks like the below:

>count_pFAM_samplelist
count pFAM sample
1 2 1A1D_ACIAC/10-323 1
2 3 1A1D_AGRRK/9-322 1
3 1 1A1D_BURCC/10-323 1
4 1 1A1D_CUPNH/9-323 1
5 1 1A1D_METNO/9-322 1
6 2 1A1D_METPP/10-323 1
7 3 1A1D_METS4/9-322 1
8 1 1A1D_PSES0/11-323 1
9 1 1A1D_PSEUD/10-323 1
10 2 1A1D_RHIRD/10-322 1
11 2 14312_ARATH/10-245 2
12 1 1433_EIMTE/9-256 2
13 1 1433_SPIOL/1-198 2
14 1 1A1D_ACIAC/10-323 2
15 4 1A1D_AGRRK/9-322 2
16 1 1A1D_CUPNH/9-323 2
17 2 1A1D_METNO/9-322 2
18 6 1A1D_METPP/10-323 2
19 1 1A1D_METS4/9-322 2
20 2 1A1D_PSEUD/10-323 2

what I would like to do is transform this into a matrix with "sample" in columns, pFAM in rows, with count listed in the matrix. If anyone has any suggestions, they would be helpful. Thanks,

-Tony

[dpryan]

Do each of the samples at least have a row for each of the pFAM IDs (assuming the example is sorted, it would seem not)? That would make life easier (and make things more amenable to conversion in R). If not, presumably you would want a 0 used when an entry is missing entirely. The general idea would be to split() that by "sample", lapply() a function to simple return the counts, and then convert that list into a matrix. But that'll only work for well behaved data (otherwise, just script something in python or whatever other language you know, since iteration is really slow in R).