SSPACE-LongRead, PBjelly2 and OPERA-LG comparison

Question

[gabri]

Hi All,

I have both a PacBio assembly and an Illumina assembly (obtained separately). I would like to perform a scaffolding - merging - of the two assemblies.

Right now, I'm trying SSPACE-LongRead (with -t 8 parameter) but it is at the finishing step from about a week. Is it normal? Did someone else have similar experiences with SSPACE-LongRead?

Anyway, I would also like to try PBjelly2 and OPERA-LG. Would you recommend these two software to do this kind of scaffolding process? Are they more suitable than SSPACE-LongRead?

Thank you very much

[Tm]

Hello, have you got answer to this question from anywhere? I am also trying to analyse same type of data and want to know which tool can give better result.

Answer 1

[AK]

Hi gabri,

You have: (1) a PacBio assembly and (2) an Illumina assembly, which sounds more suitable to use tools such as quickmerge to reconcile the two assemblies you have.

If you are interested in scaffolding your Illumina assembly by your PacBio reads, my experiences using SSPACE-LongRead or OPERA-LG are:

• You can tweak SSPACE-LongRead to use minimap2 instead of blasr, which shortens the runtime drastically;
• Consider using the corrected PacBio reads for scaffolding (for example CoLoRMap), in this way the mapping quality improves;
• Did you succeed using OPERA-LG? Not sure if it's by design but when I tried OPERA-LG using the long read version (v2.0.6, OPERA-long-read.pl), it always prompts out to ask for mate-pair Illumina library. If you have long reads only perhaps it's not the way to go.

Hope this help.