Adding metadata to genomicranges list

Question: Adding metadata to genomicranges list

6.1 years ago by

United States

reviewer3 • 20 wrote:

Dear BioStars,

I have a genomic ranges (GR) list object called "lgr.test" and would like to simply add several columns to its metadata. How would I do that? I have a list of numeric vectors "lvec.test" of the same lengths as the GRs in lgr.test.

Thank you!

> summary(lvec.test)

      Length Class  Mode   
chr1  589    -none- numeric
chr2  790    -none- numeric
chr3  482    -none- numeric
chr4  681    -none- numeric
chr5  698    -none- numeric
chr6  492    -none- numeric
chr7  713    -none- numeric
chr8  489    -none- numeric
chr9  590    -none- numeric
chr10 521    -none- numeric
chr11 853    -none- numeric
chr12 395    -none- numeric
chr13 373    -none- numeric
chr14 358    -none- numeric
chr15 388    -none- numeric
chr16 340    -none- numeric
chr17 465    -none- numeric
chr18 287    -none- numeric
chr19 357    -none- numeric

> summary(lgr.test)

      Length Class   Mode
chr1  589    GRanges S4  
chr2  790    GRanges S4  
chr3  482    GRanges S4  
chr4  681    GRanges S4  
chr5  698    GRanges S4  
chr6  492    GRanges S4  
chr7  713    GRanges S4  
chr8  489    GRanges S4  
chr9  590    GRanges S4  
chr10 521    GRanges S4  
chr11 853    GRanges S4  
chr12 395    GRanges S4  
chr13 373    GRanges S4  
chr14 358    GRanges S4  
chr15 388    GRanges S4  
chr16 340    GRanges S4  
chr17 465    GRanges S4  
chr18 287    GRanges S4  
chr19 357    GRanges S4

genomicranges genomic ranges gr list • 2.5k views

ADD COMMENT • link •

modified 6.1 years ago • written 6.1 years ago by reviewer3 • 20

6.1 years ago by

reviewer3 • 20

United States

reviewer3 • 20 wrote:

Ok, a for loop works (see below), but if some GR gurus see this, please save me from my ignorance and tell me how to do this using standard GR functions (Map , endoapply, mendoapply, Reduce). The for solution is fast enough but doesn't use the names of the objects in the lists, so it depends entirely on their order.

Cheers!

For solution:

for (i in 1:length(lgr.test)) { values(lgr.test[[i]])$NEW<-lvec.test[[i]] }

ADD COMMENT • link modified 6.1 years ago • written 6.1 years ago by reviewer3 • 20

6.1 years ago by

Philipp Bayer ♦ 6.7k

Australia/Perth/UWA

Philipp Bayer ♦ 6.7k wrote:

One lazy way is to use the merge function:

merged <- merged(lvec.test, lgr.test, by='row.names')

This won't give you the final product, it will give you something like:

Row.names Length.x Class.x Mode.x Length.y Class.y Mode.y

chr1 589 NA numeric 589 GRanges S4 ...

You can then clean it up and drop the columns you don't like, for example by

rownames(merged) <- merged[,1]

merged <- merged[-1]

merged <- merged[-2] etc.

ADD COMMENT • link modified 6.1 years ago • written 6.1 years ago by Philipp Bayer ♦ 6.7k

Hi Philipp, thanks for your answer. I tried merge but got an error. I think I have a solution using a for loop, but I'm pretty sure there's a better way to do it using Map, except that my vectors are regular numeric vectors not S4 GR objects. I'll keep hacking at it, but if you have another idea, I'll appreciate it. Thanks!

ADD REPLY • link written 6.1 years ago by reviewer3 • 20

No worries! What's the error exactly?

You may be able to circumvent it by putting as.data.frame() around lvec.test and lgr.test

ADD REPLY • link written 6.1 years ago by Philipp Bayer ♦ 6.7k

Hi Philipp, I tried both (with and without as.data.frame but get an error referring to a difference in row numbers):

arguments imply differing number of rows: 589, 790, 482, 681, 698, 492, 713, 489, 590, 521, 853, 395, 373, 358, 388, 340, 465, 287, 357

ADD REPLY • link written 6.1 years ago by reviewer3 • 20

Aaah your two tables have different row-lengths - you can try adding all.x = TRUE or all.y = TRUE to the merge command:

If all.x is true, all the non matching cases of x are appended to the result as well, with NA filled in the corresponding columns of y; analogously for all.y.

ADD REPLY • link written 6.1 years ago by Philipp Bayer ♦ 6.7k

Please log in to add an answer.

Similar posts • Search »