new features

[lpryszcz]

• AUTOMATISATION
  • estimate read limit
  • refine insert size estimation for each iteration
  • include optional genome assembly step, is it really needed?
  • include --resume option
  • installator (to assist less experienced users)
  • INSTALL.sh: UNIX installer
  • Docker image
  • Alpine Linux: strip numpy and biopython
  • official conda package bioconda package for redundans #27 unofficial one
• PERFORMANCE
  • approximate aligner ie minimap Redundans hanging on scaffold.longread stage #60
  • sorting .psl files in memory
  • memory-optimised
• SENSITIVITY
  • better similarity search algorithm ie. LAST
  • include depth of coverage
  • calculate insert size on major orientation only
• Features
  • Support for long reads (PacBio, Nanopore)
  • Logging to file
  • FastaIndex
  • fasta2homozygous.py: report removed contigs
• Exceptions
  • check LAST version
  • no read aligned due to very fragmented assembly
• Long-read scaffolding
• production version of pyScaf Lost sequences when scaffolding using long reads #51
• excessive gaps when scaffolding pacbio reads
• Code
• Python3
• remove nonsense reduction step overlap #81
• make sure single TMP is used across the pipeline reduction step overlap #81
• Output
• gfa support (low priority) Is it possible to get a gfa #72

[lpryszcz]

automatically estimate read limit setting using homozygous assembly size after reduction
and assuming let's say 50bp reads

[mictadlo]

• Miss-assembly detection: MisSEQuel
• GMcloser

[margaretc-ho]

Is there any plan to fix the excessive gaps when scaffolding pacbio reads? I couldn't get redundans to gap close with the long reads, and ended up running MASURCA's samba close_scaffold_gaps.sh with my long reads in order to get the gaps closed. It took me a while to find the right tool to do this (had problems running other gap filling tools like Gapless, GAPPadder, and MindTheGap, none of which were satisfactory). Putting a note here in case it helps others.

[lpryszcz]

hi, we decided to focus redundans on short-read technologies, so we won't be adding any long-read related features in the future.