sra toolkit does not contain a utility for reconstitution of HDF5 from PacBio SRA archives - PacBio was only concerned with fastq-dump when they contacted SRA about processing their HDF5 . there is no information lost, so you could make HDF5 from the ouput of vdb-dump. but if you want consensus, it is stored in sra table called consensus. you can get it this way: fastq-dump --table CONSENSUS SRR515631 if you have questions about sra toolkit functionality, don't hesitate to email sra@ncbi.nlm.nih.gov

-Chris

Tucked in the blasr source (https://github.com/PacificBiosciences/blasr ) is a program pls2fasta. The usage of the program is printed below.

-mark

[mchaisso@e31 test_region_assembly]$ ~/software/blasr/pbihdfutils/bin/pls2fasta 
pls2fasta Converts bas.h5 files to fasta or fastq files. Although fasta files 
         are provided with every run, they are not trimmed nor split into 
         subreads. This program takes additional annotation information, such 
         as the subread coordinates and high quality regions and uses them to 
         create fasta sequences that are substrings of all bases called. Most 
         of the time you will want to trim low quality reads, so you should 
         specify -trimByRegion.

usage: pls2fasta file.pls.h5 out.fasta [options] 

   file.pls.h5     
               Input pls/bas.h5 file.
   out.fasta       
               Output fasta/fastq file.
  -trimByRegion   
               Trim away low quality regions.
  -maskByRegion   
               Mask low quality regions with 'N'.
  -regionTable value 
               Optional HDF file with a /PulseData/Regions dataset.
  -minSubreadLength value 
               Do not write subreads less than the specified length.
  -noSplitSubreads
               Do not split reads on adapter sequences.
  -holeNumber     
               Only print this hole number (or list of numbers).
  -fastq          
               Print in FASTQ format with quality.
  -ccs            
               Print de novo CCS sequences
  -passesInTitle  
               Append /N_passes/ to ccs sequence title, where N is the number 
               of passes.
  -lineLength value 
               Specify fasta/fastq line length
  -minReadScore value 
               Minimum read score to print a read.  The score is a number 
               between 0 and 1000 and represents the expected accuracy 
               percentage * 10. A typical value would be between 750 and 800.  
               This does not apply to ccs reads.
  -best           
               If a CCS sequence exists, print this.  Otherwise, print the 
               longestsubread.  This does not support fastq.