Hello people,
Has anyone in here used OPERA-LG or PBJelly to improve assembly by better scaffolding and filling in gaps ?
Actually I am seeking explanation to few "abnormalities" that I am not able to understand:
After running OPERA-LG (Paired end data + pacbio data), the max scaffold size increased and number of scaffolds decreased...but along with that neither the genome size increased nor the gaps were removed (rather they increased !!) with regard to the PE data assembly...what can possibly be the reason ??
With PBJelly run, first blasr never run from within the script and second even if all the stages execute succesfully, in the fifth and sixth stage it refuses to assemble the sequences at all (I can see all the files being present though !!!). I have manually checked that there are actually pacbio reads spanning the gaps in the scaffolds from PE data assembly by velvet.
Any help, opinion and suggestion is highly appreciated and thanks in advance.
I don't see an abnormality in what OPERA-LG is doing. A scaffolder is supposed to decrease the number of scaffolds and increase the size but why would you expect the genome size to increase. Since you say the number of gaps increased, there were gaps already present so obviously there was another scaffolder used earlier. This wouldn't change the data much. Earlier scaffolders are already making an estimate based on the number and size of gaps that should be allowed based on insert sizes and lengths of long reads - doesn't this mean the number of gaps might increase due to more information without change in genome size estimate.
PBJelly log file is needed to comment on this. Probably your parameters are too strict to allow more scaffolding, such as the number of reads supporting? The pacbio community is active on github, probably they can give a better feedback if you provide more information. Here is a direct link to the pbjelly discussion page -
https://sourceforge.net/p/pb-jelly/discussion/general/
You need to check the version of blasr