Overview

Quake is a package to correct substitution sequencing errors in experiments with deep coverage (e.g. >15X), specifically intended for Illumina sequencing reads. Quake adopts the k-mer error correction framework, first introduced by the EULER genome assembly package. Unlike EULER and similar progams, Quake utilizes a robust mixture model of erroneous and genuine k-mer distributions to determine where errors are located. Then Quake uses read quality values and learns the nucleotide to nucleotide error rates to determine what types of errors are most likely. This leads to more corrections and greater accuracy, especially with respect to avoiding mis-corrections, which create false sequence unsimilar to anything in the original genome sequence from which the read was taken.

News

8/17/2011 - Version 0.3

Added automatic support for Jellyfish k-mer counting and made it the default mode, as well as some small improvements.

12/12/2010 - Version 0.2

Fixed a bug that produced erroneous results for paired end reads under the trimmed read minimum length threshold.

11/29/2010 - Manuscript publication

The Quake manuscript was published online today in Genome Biology.

11/23/2010 - K-mer counting

An excellent new k-mer counting program called JELLYFISH developed by Guillaume Marçais and Carl Kingsford is now available for use. JELLYFISH exploits the "compare-and-swap" CPU instruction to efficiently parallelize hash table construction. This program is a great option for Quake users who are uninterested in Hadoop and have large datasets to correct.

07/21/2010 - Version 0.1

In addition to a number of minor tweaks, the semi-significant improvements include:

  • The -u option now outputs a file of uncorrected reads, and maintains read pairs. See manual section 4 for details.
  • The -z option is now more efficient about zipping output files by using the Gzstream library.
  • For each reads file A.fastq, a summary statistics file A.stats.txt is output automatically.

07/04/2010 - Initial release

The first public release of Quake is now available for download. See the manual for instructions on how to set up and run the programs.