Denovo Hybrid Assembly using 454/illumina

[intikhab]

Hi,

I have managed to get a reasonable assembly of a bacterial genome using newbler based on 454 NG sequence data.

I have also available paired end reads for the same bacterial genome and I am wondering is there a way to improve the contigs and scaffolding.

I read about AMOS-hybrid, http://www.biomedcentral.com/1471-2164/11/242, but it requires mates file, which I do not have available instead I have one sequence file for each of the paired ends.

I also thought to try mira but it also requires mates files. Should I request the vender for mates file or there are other software available that can use contigs from my newbler run and raw paired end files and their qualities?

I hope other people may have gone through hybrid assembly issue already and may be of help.

Regards,

Intikhab

[andreas.sjodin]

I would refer you to the BAMBUS manual where you find a description of the mates format:

http://sourceforge.net/apps/mediawiki/amos/index.php?title=Bambus_Manual#The_.mates_file

You may also find an example in the test folder of AMOS-Hybrid.

library fiveK 4000 6000 (r).*
pair (.*)\.1 (.*)\.2

The first row describes the library information and the second row contains mate-pair relationships. You may create your own mate-file based on information your own library and read structure.

[themerlin]

Hi,

I have managed to get a reasonable assembly of a bacterial genome using newbler based on 454 NG sequence data.

I have also available paired end reads for the same bacterial genome and I am wondering is there a way to improve the contigs and scaffolding.

If you wanted to use Mira, there are instructions on how to directly use Mira with Bambus:

http://mira-assembler.sourceforge.net/docs/scaffolding_MIRA_BAMBUS.pdf

[intikhab]

About the mates file, if I have the following description of the paired ends reads:

>NG-5247_HLP_3_1_1058_5238/1
ATGATGCATCTGAGACGATNGGTGTAAACGATCGT

>NG-5247_HLP_3_1_1058_5238/2
ACAAGATATATACGTATTATTAGGGTCAATAATGG

How should the mates file look a like?:

library NG-5247 150 200 (^NG).*
pair (.*)\/1 (.*)\/2


One related question. When we have two separate files for paired end reads, how can one provide these to mira or AMOS-Hybrid, these programs require one file. Does, jut combining the two e.g. using cat a b >c should do?

Intikhab

I would refer you to the BAMBUS manual where you find a description of the mates format:

http://sourceforge.net/apps/mediawiki/amos/index.php?title=Bambus_Manual#The_.mates_file

You may also find an example in the test folder of AMOS-Hybrid.


The first row describes the library information and the second row contains mate-pair relationships. You may create your own mate-file based on information your own library and read structure.

[sbberes]

I suggest checking MIRA again, I think that you are in error in thinking that it requires mates files. I have used MIRA with good results doing hybrid assemblies of single end read 454 and illumina data. See the thread: "Combining 454FLX and SOLiD runs for de novo genome assembly", I outlined the process that I used there.
SBB

[intikhab]

Here I am using AMOS-Hybrid, where mates file is required. For mira I know we dont need mates file, what we need there is a caf file from an initial assembly and sequence file from the second technology.

Can anybody correct me on the mates file, mentioned above. When I use this, the error log shows a large number of reads got excluded.

I want to know specifically the regular expression for the library and for the forward and reverse reads, I have mentioned the description of forward and reverse reads above.

Regards,

Intikhab