de novo assembly of repeat elements

HESmith · 09-22-2011, 05:36 PM

As part of a de novo assembly project, I'd like to try to identify repeat elements - everything from single gene duplications (difficult) to transposons (less difficult). The data are Illumina PE-101 reads, ~50X coverage. My (admittedly unsophisticated) approach is to assemble contigs (I'll try both de Bruijn and overlap assemblers), then flag those with >2X average read depth.

Two questions:
1) are there any tools designed for this application?
2) any suggestions for alternative strategies (e.g., candidate identification by sequence conservation, branch counting of de Bruijn graphs, etc.)?

Thanks,
Harold

jimmybee · 09-22-2011, 11:04 PM

Whats your species?

Hobbe · 09-23-2011, 07:00 AM

Repeatmasker both identifies and masks repeat-elements. Some assembly programs, like Mira, also mark repeat regions with tags. In Mira you can check the sequences identified as repeats in the projectname_info_readrepeats.lst file.

HESmith · 09-23-2011, 09:03 AM

Quote:

Originally Posted by jimmybee

Whats your species?

Nematodes for now, but there are likely to be others in the future.

Harold

HESmith · 09-23-2011, 09:03 AM

Thanks for the recommendations, Hobbe. I'll look into them.

Harold

saemi · 10-26-2011, 03:48 PM

Hi Harold

How is your project regarding the de novo assembly of transposons going? I'm interested in doing a similar project to compare transposons among closely related plant species using illumina sequencing. What programs are you using?

Cheers,Saemi

pmiguel · 10-27-2011, 06:19 AM

For plants you will see lots of LTR retrotransposons. These might assemble with the LTRs in the middle -- since the LTRs are, well long (0.2-5 kb, for the most part) and also, repeats that flank the internal domains of these ubiquitous transposable elements.

There was even an program designed by Jeremy DeBarry when he was at UGA in Bennetzen lab, that took advantage of this to pull LTR retros from full genome assemblies and reconstruct their LTRs in the correct positions. He called it the "AAARF" algorithm. (Get it? UGA Bulldogs, aaarf?)

Ah, here it is. Also a publication. Looks like it is also usable for other sorts of elements as well. There you go: code from a maize lab -- you know maize, the organism where transposable elements were discovered? Worth a look.

--
Phillip

HESmith · 10-27-2011, 07:45 AM

Hi Saemi,

I'm still waiting to obtain the sequence data for this project, so I don't have any results to report. I'll keep you posted regarding my progress.

Harold

saemi · 10-27-2011, 12:43 PM

Hi

@Harold, OK great, I'm looking forward to hear from you.

@Phillip, Great thank you very much for the information and the paper. I'll will take a close look at it. One of the things I'm concerned about is the fact that I plan to use an Illumina Hi-Seq in my project, on species which don't have a reference genome. Most of the available methods for looking at transposons in a shotgun library I've seen, work on 454 sequences. I guess one way to go is to do a de novo assembly first on the data but them I'm afraid to loose information from my dataset.

Thank you guys
Saemi

Claudia34 · 11-09-2011, 03:03 AM

Hi,

We have the project to identify genome-wide transposition events in flies after know-down of a protein of interest. We think sequencing and de novo assembly are the best way to do it. Do you agree ?
I don't know if Illumina is the best technology for these kinds of analysis because of the sequencing length. Do you have any recommendation?

Thanks,
Claudia

HESmith · 11-09-2011, 07:38 AM

Hi Claudia,

If you already have a genome assembly for your species and you know the sequences of your transposons, you can use the strategy described here. Briefly, use paired-end sequencing and map the different ends to genomic and transposon sequences to identify insertion sites. Let me know if you want more details.

Harold

Claudia34 · 11-09-2011, 07:50 AM

Hi Harold,

Thanks a lot for your answer. I think we can use this strategy because we are working on drosophila melanogaster. I will carefully read this paper.

Thanks again,
Claudia

adaptivegenome · 11-09-2011, 04:11 PM

Quote:

Originally Posted by Claudia34

Hi Harold,

Thanks a lot for your answer. I think we can use this strategy because we are working on drosophila melanogaster. I will carefully read this paper.

Thanks again,
Claudia

What are you suppressing in the flies, Hsp90?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
De Novo Assembly using Ray	Farhat	De novo discovery	18	05-23-2012 02:19 PM
How to get started in de novo assembly?	ymc	Bioinformatics	4	01-14-2012 12:29 AM
how to resolve repeat areas with Velvet when doing de novo assembly	salmonella	De novo discovery	1	10-24-2011 09:42 PM
De novo assembly	mihir.karnik	General	1	09-07-2011 02:49 PM
de novo assembly vs. reference assembly	fadista	General	3	02-16-2011 12:11 AM

09-22-2011, 05:36 PM	#1
HESmith Senior Member Location: Bethesda MD Join Date: Oct 2009 Posts: 509	de novo assembly of repeat elements As part of a de novo assembly project, I'd like to try to identify repeat elements - everything from single gene duplications (difficult) to transposons (less difficult). The data are Illumina PE-101 reads, ~50X coverage. My (admittedly unsophisticated) approach is to assemble contigs (I'll try both de Bruijn and overlap assemblers), then flag those with >2X average read depth. Two questions: 1) are there any tools designed for this application? 2) any suggestions for alternative strategies (e.g., candidate identification by sequence conservation, branch counting of de Bruijn graphs, etc.)? Thanks, Harold

09-22-2011, 11:04 PM	#2
jimmybee Senior Member Location: Adelaide, Australia Join Date: Sep 2010 Posts: 119	Whats your species?

09-23-2011, 07:00 AM	#3
Hobbe Member Location: Uppsala, Sweden Join Date: Apr 2010 Posts: 29	Repeatmasker both identifies and masks repeat-elements. Some assembly programs, like Mira, also mark repeat regions with tags. In Mira you can check the sequences identified as repeats in the projectname_info_readrepeats.lst file.

09-23-2011, 09:03 AM	#5
HESmith Senior Member Location: Bethesda MD Join Date: Oct 2009 Posts: 509	Thanks for the recommendations, Hobbe. I'll look into them. Harold

10-26-2011, 03:48 PM	#6
saemi Junior Member Location: University of British Columbia, Vancouver Join Date: Oct 2010 Posts: 5	Hi Harold How is your project regarding the de novo assembly of transposons going? I'm interested in doing a similar project to compare transposons among closely related plant species using illumina sequencing. What programs are you using? Cheers,Saemi

10-27-2011, 06:19 AM	#7
pmiguel Senior Member Location: Purdue University, West Lafayette, Indiana Join Date: Aug 2008 Posts: 2,317	For plants you will see lots of LTR retrotransposons. These might assemble with the LTRs in the middle -- since the LTRs are, well long (0.2-5 kb, for the most part) and also, repeats that flank the internal domains of these ubiquitous transposable elements. There was even an program designed by Jeremy DeBarry when he was at UGA in Bennetzen lab, that took advantage of this to pull LTR retros from full genome assemblies and reconstruct their LTRs in the correct positions. He called it the "AAARF" algorithm. (Get it? UGA Bulldogs, aaarf?) Ah, here it is. Also a publication. Looks like it is also usable for other sorts of elements as well. There you go: code from a maize lab -- you know maize, the organism where transposable elements were discovered? Worth a look. -- Phillip

10-27-2011, 07:45 AM	#8
HESmith Senior Member Location: Bethesda MD Join Date: Oct 2009 Posts: 509	Hi Saemi, I'm still waiting to obtain the sequence data for this project, so I don't have any results to report. I'll keep you posted regarding my progress. Harold

10-27-2011, 12:43 PM	#9
saemi Junior Member Location: University of British Columbia, Vancouver Join Date: Oct 2010 Posts: 5	Hi @Harold, OK great, I'm looking forward to hear from you. @Phillip, Great thank you very much for the information and the paper. I'll will take a close look at it. One of the things I'm concerned about is the fact that I plan to use an Illumina Hi-Seq in my project, on species which don't have a reference genome. Most of the available methods for looking at transposons in a shotgun library I've seen, work on 454 sequences. I guess one way to go is to do a de novo assembly first on the data but them I'm afraid to loose information from my dataset. Thank you guys Saemi

11-09-2011, 03:03 AM	#10
Claudia34 Junior Member Location: Geneva, Switzerland Join Date: Sep 2010 Posts: 9	Hi, We have the project to identify genome-wide transposition events in flies after know-down of a protein of interest. We think sequencing and de novo assembly are the best way to do it. Do you agree ? I don't know if Illumina is the best technology for these kinds of analysis because of the sequencing length. Do you have any recommendation? Thanks, Claudia

11-09-2011, 07:38 AM	#11
HESmith Senior Member Location: Bethesda MD Join Date: Oct 2009 Posts: 509	Hi Claudia, If you already have a genome assembly for your species and you know the sequences of your transposons, you can use the strategy described here. Briefly, use paired-end sequencing and map the different ends to genomic and transposon sequences to identify insertion sites. Let me know if you want more details. Harold

11-09-2011, 07:50 AM	#12
Claudia34 Junior Member Location: Geneva, Switzerland Join Date: Sep 2010 Posts: 9	Hi Harold, Thanks a lot for your answer. I think we can use this strategy because we are working on drosophila melanogaster. I will carefully read this paper. Thanks again, Claudia