ReptileErrorCorrection

Introduction

This page will describe how to use NGOpt tools with Reptile to correct errors in Illumina reads.

Prerequisites

You will need the following packages/programs. They will need to be in your executable path:

• Reptile - error-correction software
• NGOpt toolkit

Step 1. Convert Fastq to Fasta+Qual Files.

In the directory containing my_reads.fastq, run:

  <ngopt_trunk>/tools/quality/reptile_fq-converter ./ ./  2

This will create three files, my_reads.fa , my_reads.q, and my_reads.map.txt. Because Reptile changes the original sequence names to arbitrary names, I have posted a hack on NGOpt that will produce a .map.txt file. We will use this file later to rename our reads to the original names.

Note: All files ending in .fastq in the current directory will be processed.

Step 2. Create Reptile Config-File

Copy the Reptile config-file template from the Reptile directory and change as needed.

  cp <path_to_reptile>/src/config~ ./my_reads.rec.conf
  vim my_reads.rec.conf

Edit my_reads.rec.conf to specify input and output files.

• Modify line beginning with "InFaFile" to contain my_reads.fa
• Modify line beginning with "IQFile" to contain my_reads.q
• Modify line beginning with "OErrFile" to contain the desired name of the output file.
  • This output file will not be corrected reads. We are going to run another command on this file to get our corrected reads. Let's call the file reptile-output
    If you are more educated in the Reptile Error Correction process, feel free to modify any of other run configurations.

Step 3. Run Reptile Error Correction

Run reptile with the config file from previous step.

  reptile my_reads.rec.conf

This will produce the file reptile-output, as mentioned in the previous step.

Step 4. Merge Corrections

Now that we've identified our errors, we will merge them to get back corrected reads.

  reptile_merger my_reads.fa reptile-output my_reads.rec.fasta

The third argument to reptile_merger is the output file. This file will contain our corrected reads.

Step 5. Rename Corrected Reads

To get back the original read names, we can use a script from NGOpt--reptile_rename.

<ngopt_trunk>/tools/quality/reptile_rename my_reads.map.txt my_reads.rec.fasta > my_reads.rec.renamed.fasta

We now have corrected reads with their original names in my_reads.rec.renamed.fasta.

Where to next?

If you have a paired-end library, you might want to re-pair reads with broken pairings. You can use the program repair in the NGOpt toolkit to fix broken pairings.

repair -p ./ -s .fasta my_reads.rec my_reads.rec.renamed.fasta

This will produce three files: my_reads.rec_p1.fasta, my_reads.rec_p2.fasta, and my_reads.rec_up.fasta. The third file contains reads with a missing pair.

[Anonymous]

Originally posted by: eks.y...@gmail.com

Reptile Error correction has been updated to include a tool for choosing default parameters for each input dataset