Skip to content

Latest commit

 

History

History

docs

Table of Contents

Redundans docs

FAQ

Estimation of my library statistics is incorrect. Can I specify these values manually?

This can happen for highly fragmented assemblies or poor quality libraries. You can specify library statistics manually. To do so, look for *.is.txt file specific for your library i.e. for -i 5000_1.fq.gz 5000_2.fq.gz you will have to enter requested values into 5000_2.fq.gz.is.txt. .is.txt file is tab-delimited, with following columns:

  • read length
  • median insert size
  • mean insert size
  • standard deviation of insert size
  • and number of reads with FF, FR, RF & RR orientation.

Make sure you specify some large number of mates for requested orientation (ie 100,000). For example if you want mate-pairs with RF orientation and 5kb insert size +/- 1.5kb, enter into respective *.is.txt file:

60  5000.0        5000.0  1500.0  0       0      100000   0

How is multiple redundancy handled?

Redundans removes all contigs, but the longest one, that fullfill identity & overlap critaria during reduction step. For more info see issue #8.

Why there are two github repositories for Redundans?

https://github.com/Gabaldonlab/redundans is the official repository for Redundans, but we keep https://github.com/lpryszcz/redundans for back-compatibility, as some of the very first users of Redundans use it. Both Redundans repositories contain the same code and are regularly updated.

How to cite Redundans?

Leszek P. Pryszcz and Toni Gabaldón (2016) Redundans: an assembly pipeline for highly heterozygous genomes. NAR. doi: 10.1093/nar/gkw294

FAQ - INSTALL.sh