How to build combined library for repeatmasker?

Shishir · 02-11-2013, 03:16 AM

I want to to perform repeatmasking on ant genome. Ant is also included in the repeat library obtained from Repbase. The library RepeatMaskerLib.embl also contains my species. However, I also generated de novo repeat library by repeatmodeler (consensi.fa). How can I use both library for masking ant genome ? Should I concatenate RepeatMaskerLib.embl and consensi.fa or can I use both library with single command?

suryasaha · 02-11-2013, 06:21 AM

You can go both ways. Have not done this myself but the simplest way would be to pull put ant repeats from RepeatMaskerLib.embl and merge them with your custom repeat library (as fasta). You could then use the -lib option in RepeatMasker to supply the updated custom repeat library and mask the ant genome.

AFAIK, you cannot combine both libraries in a single command.

HeyIamNuria · 05-29-2013, 08:22 AM

Hey, I think this would be a great solution, but I have a problem...
I want to mask the repeats, but I want to classify them as well.

If I get the Repbase library in fasta format (to concatenate it with my own database and use it with -lib) then I lose the classificators (ex: #LINE/L1 --> ).

[I don't know if anyone came out with a script to put the "#" and "/" where they should be, but I don't know how, because usually there's only the Superfamily but not the Order (ex L1 but not LINE)]

I also tried to convert RepeatMaskerLib.embl to .fasta but I lose the classification info too.

I guess it won't work to put my repeats, in .embl, into RepeatMaskerLib.embl, but I haven't tried yet.

Did anyone find a way to solve this?
Thank you for any help

HeyIamNuria · 05-30-2013, 04:04 AM

or maybe...

Does anyone know how to use RepeatClassifier (part of RepeatModeler) only. Or make RepeatModeler classify a bunch of repeats already listed?

Thanks

rhubley · 06-11-2013, 11:41 AM

RepeatClassifier can be run independently. If you hand it a fasta file it will generate a new file with a *.classified suffix. The output file will be a fasta file with all the id#class/subclass identifiers.

rhubley · 06-11-2013, 01:08 PM

Check out the util/buildRMLibFromEMBL.pl script included in the RepeatMasker distribution. This will convert the RepeatMaskerLib.embl file into a fasta file with the "#class/subclass" nomenclature.

HeyIamNuria · 06-14-2013, 07:05 AM

Thank you for your help.
I manage to get my library classified, but I still have another doubt.

When I use Repeatmasker with a custom library (-lib) the .tbl file seems to be made with the results of masking the genome with
Homo sapies repeats alone ("The query species was assumed to be homo"). My guess is that I can't obtain a .tbl output with a custom library (even if this library is made by RepeatModeler). Is this true?

Thank you again

Nuria

HeyIamNuria · 06-28-2013, 02:03 AM

If the answer to my previous question is yes (so I can't get a real .tbl output when using RepeatMasker with the -lib option) is there any script that can produce something similar to the .tbl output with RepeatMasker .out or .gff files?

(My library is classified with #class/subclass identifiers).

Thank you for your time

Nuria

rhubley · 06-28-2013, 08:04 AM

Yes. You can use the RepeatMasker/util/buildSummary.pl script to reprocess the *.out file and summarize the results. The buildSummary.pl creates a similar table to the *.tbl file but also includes tabulations of each individual repeat and each seq/chr of the input.

HeyIamNuria · 07-11-2013, 04:28 AM

Thank you, buildSummary seems to be what I was looking for, but I have another question.

I masked the same genome with different libraries and I used the buildSummary script to extract the repeat information from the *.out files.

But every .tbl file considers a different genome length. It is a number between the original total length and the excluding N/X-runs value. In my case 10 to 20 Mb smaller that the original length.

Consequently the % of repeats have important differences.

Can anyone help?

Thank you very much

Nuria

HeyIamNuria · 09-25-2013, 01:37 AM

ok, I only had to read the buildSummary.pl script.

If someone else has the same problem, I solved using:
perl buildSummary.pl -useAbsoluteGenomeSize -genome tablewithscaffoldsnamesandlength.tsv -species yourspecies RMaskeroutput.out

Nuria

Pseudoknot · 03-01-2015, 07:41 PM

Hi Nuria,

I'm having the same problem. Just to be clear, you masked your genome using the RepeatModeler results and then used only the buildSummary.pl script to classify the repeats?

Thanks!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
question about Repeatmasker	flyyuan	Bioinformatics	14	04-10-2014 03:04 PM
How to construct a combined library for repeatmasker	sunhh	Bioinformatics	17	02-12-2014 08:59 PM
RepeatMasker error	parulvk	Bioinformatics	0	10-26-2011 02:59 PM
RepeatMasker output	flipwell	Bioinformatics	5	08-17-2011 04:48 PM
RepeatMasker question	slny	Bioinformatics	0	06-08-2011 01:31 PM

02-11-2013, 03:16 AM	#1
Shishir Member Location: Germany Join Date: Nov 2012 Posts: 22	How to build combined library for repeatmasker? I want to to perform repeatmasking on ant genome. Ant is also included in the repeat library obtained from Repbase. The library RepeatMaskerLib.embl also contains my species. However, I also generated de novo repeat library by repeatmodeler (consensi.fa). How can I use both library for masking ant genome ? Should I concatenate RepeatMaskerLib.embl and consensi.fa or can I use both library with single command?

06-14-2013, 07:05 AM	#7
HeyIamNuria Member Location: Barcelona Join Date: Dec 2012 Posts: 19	Thank you for your help. I manage to get my library classified, but I still have another doubt. When I use Repeatmasker with a custom library (-lib) the .tbl file seems to be made with the results of masking the genome with Homo sapies repeats alone ("The query species was assumed to be homo"). My guess is that I can't obtain a .tbl output with a custom library (even if this library is made by RepeatModeler). Is this true? Thank you again Nuria Last edited by HeyIamNuria; 06-21-2013 at 07:18 AM. Reason: The problem was not clearly explained.

02-11-2013, 06:21 AM	#2
suryasaha Member Location: New York Join Date: Mar 2011 Posts: 27	You can go both ways. Have not done this myself but the simplest way would be to pull put ant repeats from RepeatMaskerLib.embl and merge them with your custom repeat library (as fasta). You could then use the -lib option in RepeatMasker to supply the updated custom repeat library and mask the ant genome. AFAIK, you cannot combine both libraries in a single command.

05-29-2013, 08:22 AM	#3
HeyIamNuria Member Location: Barcelona Join Date: Dec 2012 Posts: 19	Hey, I think this would be a great solution, but I have a problem... I want to mask the repeats, but I want to classify them as well. If I get the Repbase library in fasta format (to concatenate it with my own database and use it with -lib) then I lose the classificators (ex: #LINE/L1 --> ). [I don't know if anyone came out with a script to put the "#" and "/" where they should be, but I don't know how, because usually there's only the Superfamily but not the Order (ex L1 but not LINE)] I also tried to convert RepeatMaskerLib.embl to .fasta but I lose the classification info too. I guess it won't work to put my repeats, in .embl, into RepeatMaskerLib.embl, but I haven't tried yet. Did anyone find a way to solve this? Thank you for any help

05-30-2013, 04:04 AM	#4
HeyIamNuria Member Location: Barcelona Join Date: Dec 2012 Posts: 19	or maybe... Does anyone know how to use RepeatClassifier (part of RepeatModeler) only. Or make RepeatModeler classify a bunch of repeats already listed? Thanks

06-11-2013, 11:41 AM	#5
rhubley Member Location: Washington Join Date: Sep 2012 Posts: 10	RepeatClassifier can be run independently. If you hand it a fasta file it will generate a new file with a *.classified suffix. The output file will be a fasta file with all the id#class/subclass identifiers.

06-11-2013, 01:08 PM	#6
rhubley Member Location: Washington Join Date: Sep 2012 Posts: 10	Check out the util/buildRMLibFromEMBL.pl script included in the RepeatMasker distribution. This will convert the RepeatMaskerLib.embl file into a fasta file with the "#class/subclass" nomenclature.

06-28-2013, 02:03 AM	#8
HeyIamNuria Member Location: Barcelona Join Date: Dec 2012 Posts: 19	If the answer to my previous question is yes (so I can't get a real .tbl output when using RepeatMasker with the -lib option) is there any script that can produce something similar to the .tbl output with RepeatMasker .out or .gff files? (My library is classified with #class/subclass identifiers). Thank you for your time Nuria

06-28-2013, 08:04 AM	#9
rhubley Member Location: Washington Join Date: Sep 2012 Posts: 10	Yes. You can use the RepeatMasker/util/buildSummary.pl script to reprocess the .out file and summarize the results. The buildSummary.pl creates a similar table to the .tbl file but also includes tabulations of each individual repeat and each seq/chr of the input.

07-11-2013, 04:28 AM	#10
HeyIamNuria Member Location: Barcelona Join Date: Dec 2012 Posts: 19	Thank you, buildSummary seems to be what I was looking for, but I have another question. I masked the same genome with different libraries and I used the buildSummary script to extract the repeat information from the *.out files. But every .tbl file considers a different genome length. It is a number between the original total length and the excluding N/X-runs value. In my case 10 to 20 Mb smaller that the original length. Consequently the % of repeats have important differences. Can anyone help? Thank you very much Nuria

09-25-2013, 01:37 AM	#11
HeyIamNuria Member Location: Barcelona Join Date: Dec 2012 Posts: 19	ok, I only had to read the buildSummary.pl script. If someone else has the same problem, I solved using: perl buildSummary.pl -useAbsoluteGenomeSize -genome tablewithscaffoldsnamesandlength.tsv -species yourspecies RMaskeroutput.out Nuria

03-01-2015, 07:41 PM	#12
Pseudoknot Junior Member Location: Australia Join Date: May 2012 Posts: 4	Hi Nuria, I'm having the same problem. Just to be clear, you masked your genome using the RepeatModeler results and then used only the buildSummary.pl script to classify the repeats? Thanks!