Hybrid assembling using PacBioToCA?

[metheuse]

This may be a silly question. I know PacBioToCA can be used to correct PacBio long reads by alignment from, for example, Illumina short reads. Then the corrected PacBio reads can be assembled, for example, by Celera assembler. Is this "hybrid" assembly?
In this pipeline, illumina reads are only used for error correction, but not used in assembling? Should both PacBio reads and illumina reads be used in Celera for assembling as "hybrid" assembly?

[krobison]

Hybrid in sense that assembly is driven by two technologies.

There are all sorts of variants of hybrids -- cleaning reads (as you've described), co-assembly, merging independent assemblies made with different technologies & assembling with one technology & correcting the consensus with another -- and mix-and-match of all of those

[metheuse]

Hybrid in sense that assembly is driven by two technologies.

There are all sorts of variants of hybrids -- cleaning reads (as you've described), co-assembly, merging independent assemblies made with different technologies & assembling with one technology & correcting the consensus with another -- and mix-and-match of all of those

Thanks. So my follow-up question is: is the pacbiotoca--> celera workflow a good way of "hybrid" assembly? Does it fully use the information from two technologies? As far as I know, this is the only built-in "hybrid assembly" method in PacBio's smrtanalysis pipeline. Is it generally good?

[rhall]

PacBioToCA can work well for small assemblies, computation quickly becomes an issue with larger assemblies. Smrtanalysis also includes AHA (A Hybrid Assembler) for scaffolding, but that is also limited by assembly size.

[metheuse]

PacBioToCA can work well for small assemblies, computation quickly becomes an issue with larger assemblies. Smrtanalysis also includes AHA (A Hybrid Assembler) for scaffolding, but that is also limited by assembly size.

Thanks, Rhall! You've been always helpful to me!
This should be a small genome. I'm not familiar with AHA. Is it something one can combine with the pacbioToCA-celera pipeline, or is it a parallel pipeline?

[rhall]

AHA is a separate scaffolding pipeline, if you have an illumina / 454 assembly <20,000 contigs <160 Mbase you can use PacBio long reads and AHA to scaffold and close gaps in the assembly.
You could also assembly with PacBioToCA then try scaffolding the resulting assembly with AHA in order to recover any of the sequence that was not in the PacBioToCA assembly due to illumina coverage bias.
Results from AHA should always be validated by resequencing of raw reads, to check for missassemblies.

[krobison]

Another hybrid assembler worth noting is Cerulean

http://sourceforge.net/projects/ceruleanassembler/

and PBJelly is another PacBio scaffolder

http://sourceforge.net/p/pb-jelly/wiki/Home/

[rhall]

+1 for Cerulean, I have seen some really good results for small assemblies and it is computationally efficient.
Another option for a small genomes, even if you don't have enough coverage to get great denovo PacBio results, HGAP you will still give reasonable sized assemblies with 35-40x coverage, just very fragmented. You could assemble everything independently and merge the assemblies with something like minimus2.

[metheuse]

Thank you guys so much. I'll take a look into Cerulean.

[akorobeynikov]

SPAdes 3.0 (http://bioinf.spbau.ru/en/spades) includes support for hybrid assemblies from Illumina/IonTorrent+PacBio. Additional information (like old contigs, Sanger reads) can be added as well.