get average gene length file

[emolinari]

Hi everyone,

I have a very quick question...probably super easy for experienced programmers!
I need to generate an average gene length file from the .gtf file I downloaded from Ensembl. In detail, I will need a 2 column file with ENSG id on the first and gene length on the second column (I need for HTseq normalization...)

Can anyone help me with scripting?

Thanks!!!
Manu

[dpryan]

In R:

library(GenomicFeatures)
txdb <- makeTranscriptDbFromGFF("foo.gtf", type="gtf")
trans <- transcripts(txdb, columns=c("GENEID"))
df <- data.frame(gene=trans$GENEID, len=width(trans))

You probably need to specify something else for "GENEID". You can see available options with columns(txdb).

[emolinari]

Thank you dpryan!

I get this error message
txdb <- makeTranscriptDbFromGFF("foo.gtf", type="gtf")
Error in makeTranscriptDbFromGFF("foo.gtf", type = "gtf") :
unused argument (type = "gtf")

what is it?

[dpryan]

It should have been format= rather than type=. Mea culpa.

[emolinari]

thank you!
it's running now...i'll update the result soon!!!

[emolinari]

Hi dpryan!


after 10 minutes of running this command:

In R:

df <- data.frame(gene=trans$GENEID, len=width(trans))

R Studio simply crashes... any idea why???
It takes forever to run txdb <- makeTranscriptDbFromGFF("genes.gtf", format="gtf"), but it ultimately does it in 15 min or so...

Thanks!
Manu

[dpryan]

Presumably you're running out of memory and that's causing the crash. R isn't always the best when it comes to memory. You might quickly do:

head(trans$GENEID)
head(width(trans))

just to ensure that there's nothing strange that happened while making the "trans" object. The odds of an error there are crazy low, but it couldn't hurt to double check.

BTW, the txdb object can be saved to a file, should you ever need it again. Just use something like

saveDb(txdb, file="myOrganism.sqlite")

and you can simply load it again later with

loadDb("myOrganism.sqlite")

This ends up saving a lot of time if you work with the same organism often.

[emolinari]

Presumably you're running out of memory and that's causing the crash. R isn't always the best when it comes to memory. You might quickly do:

head(trans$GENEID)
head(width(trans))

just to ensure that there's nothing strange that happened while making the "trans" object. The odds of an error there are crazy low, but it couldn't hurt to double check.

BTW, the txdb object can be saved to a file, should you ever need it again. Just use something like

saveDb(txdb, file="myOrganism.sqlite")

and you can simply load it again later with

loadDb("myOrganism.sqlite")

This ends up saving a lot of time if you work with the same organism often.

Thanks dpryan!
as a matter of fact my computer was a bit cranky...it worked perfect after i rebooted it!!!
Manu