How To: Contig to Scaffold?

[fahmida]

Hi,

I would be grateful to have suggestions regarding the followings:

1) I have a set of contigs from the de novo assembly of ~2GB 454Flx data using the 'newbler' assembler (version 2.0.01). This version of 'newbler' do not create any scaffold from the contigs. I just wonder what tool I may use to make scaffolds using the contigs. I have the following output files from 'newbler': 454contigs.fna, 454contigs.ace and 454contigs.qual

2) For the same species, I have contigs from the assembler 'SoapDeNovo' using Illumina paired-end reads. I just have the 'Soap_contigs.fna' contig file. I have blasted 'newbler' contigs against 'SoapDeNovo' contigs. How I may use contigs from two different platforms to make scaffolds.

Both the assemblies are done outside as we don't have in-house HPC facility. For the task described above I just have my own desktop machine (Mac OSX 10.6, 2.26Ghz Core2Duo, 4GB RAM).

Thanks in advance for any help.

Regards.

[odysseus]

Hi~

1) Are contigs created from paired-end reads? If they are, you should assemble them using paired-end assembly function of Newbler. After paired-end assembly, you can get a set of scaffold.

2) In my think, if you have raw reads, you can hybrid assemble them using hybrid assembler such as MIRA3. If you solve this difficulty, please inform me about the solution.

Cheers! ^^

[francesco.vezzi]

MIRA3 is probably what you need...
Remember that SOAPdenovo gives as output a file named *.scafSeq that contains the scaffolds (from my experience the contig file contains a lot of very short and repetitive contigs often shorter than the read length!!!)

Another solution is to give the output of the two assemblers as input for a "standard" assembler like PCAP or ARACHNE...

Anyway I think that you resources are really too low...

[hugh_hang]

MIRA3 is probably what you need...
Remember that SOAPdenovo gives as output a file named *.scafSeq that contains the scaffolds (from my experience the contig file contains a lot of very short and repetitive contigs often shorter than the read length!!!)

Another solution is to give the output of the two assemblers as input for a "standard" assembler like PCAP or ARACHNE...

Anyway I think that you resources are really too low...

if my data is 454 sequencing data from cDNA, is it necessary to generate scaffolds? OR can the scaffolds make any sense?