Transcriptome hybrid assembly 454 + Illumina

[Giorgio]

Hello,



I have tons of data (about 1Tb) of RNA-Seq coming from different technologies (i.e. 454 - single end, Illumina both single and paired end) of the same non-genome annotated species from different tissues/conditions. So far for the illumina data I did several assemblies with Trinity, specifically one for every single distinct run and also a huge one that derived from all the illumina reads merged together (the latter tooks about 6TBs of storage during writing of intermediate files!!!).

Everything finished with no problem.



I have also a smaller quota of 454 data that I would like to integer into the illumina data. My question is what would you suggets to do?



I thought of:



1) Merging all the 454 data in one .fasta file and run Trinity.



2) After that merging all the Trinity.fasta files obtained from the 454, the illumina single experiments and the illumina merged one (the huge) and run again trinity (with maybe normalization).



3) Doing some downstream like cd-hit-est or cap3 to remove redundancy.



What is your suggestion?



Thanks in advance,

~Giorgio

[bastianwur]

I'd try to first error correct the 454, then merge everything together and run as one input.
But that's just my first reaction, and I haven't tried that before.

[Giorgio]

Thank you! What do you mean exactly by error correct?