GBSX: a toolkit for experimental design and demultiplexing genotyping by sequencing experiments

Overview

Genotyping by Sequencing is an emerging technology for cost effective variant discovery and genotyping. However, current analysis tools do not fulfill all experimental design and analysis needs.

GBSX is a package of tools to first aid in experimental design, including choice of enzymes and barcode design. Secondly, it provides a first analysis step to demultiplex samples using in-line barcodes, providing fastq files that can easily be plugged into existing variant analysis pipelines.

Download

The perl script for in silico digests and the compiled program for all other analyses can be found in the releases directory. The latest directory has the latest version. However, previous versions are still available. The complete source code can be found in the src directory. Example data and results for the tool can be found in the example directory.

##Licence

All parts of this tool is licenced under GPLv3.
A copy of this licence is included under LICENSE.

Contact

Genomics Core
Center for Human Genetics
UZ – KU Leuven
Herestraat 49 PO box 602
B-3000 Leuven, Belgium

Mail: koen.herten@kuleuven.be

Citing GBSX

We ask that you cite this paper if you use GBSX in work that leads to publication.

Herten, Koen and Hestand, Matthew S. and Vermeesch, Joris R. and Van Houdt, Jeroen KJ (2015) GBSX: a toolkit for experimental design and demultiplexing genotyping by sequencing experiments BMC Bioinformatics 2015, 16:73 doi:10.1186/s12859-015-0514-3

Help

Genotyping By Sequencing demultipleXing toolkit (GBSX) is a toolkit with an inline barcode demultiplexer for usage in the analysis of single read or paired-end genotyping by sequence (GBS) data, a barcode generator, a barcode discovery tool, and a restriction enzyme predictor. GBSX can easily be incorperated as a preceding analysis step for already deployed SNP pipelines.

Restriction Enzyme Predictor

mandatory parameters:

• -d digest sequence

• `-l`     read length   
  

• `-f`     file of reference fasta file location(s)  
  

optional parameters:

• -e enzyme name to use (default: Enzyme)

• `-g`     genome name to use in bed file name (default: genome)  
  

• `-n`     minimum size fragments to include (default: 100)  
  

• `-m`     maximum size fragments to use (default: 1000)  
  

• `-E`     second enzyme name to use (default: Enzyme2)  
  

• `-D`     digest sequence for a second enzyme (default: not declared)  
  

• `-R`     digest sequence for a third enzyme (default: not declared)  
  

Barcode Generator

mandatory parameters:

• -b the number of barcodes needed
• -e the enzyme used for the experiment

optional parameters:

• -ef the enzyme file. This option adds new enzymes.
  The file must be tab delimited: First column the enzyme name, second column the cutsites remains (comma separated) (example: enzyme ApeKI and restriction site G^CWGC: "ApeKI \tab CAGC,CTGC").
• -nb the maximum number of bootstraps that must be executed. (optional, standard 10000). By the start of a new bootstrap a complete new design is made. The best scored design (most random barcodes and best scored bases distribution is kept as result)
• -bt the number of barcode tries. (standard 20) If a random barcode does not fit into the current design try this number of times with a new random barcode before restarting the bootstrap.
• -o the output directory (standard current working directory)
• -us try to find the ultime match: the best barcode combination with the best bases distribution (standard false) true: continue even when the right number of barcodes is found.
• -bf a file with all barcodes that are used as basic set (this file is one of the possible output files)
• -nf a file with all barcodes that may not be used in the design. If this file contains barcodes that are also found in the basic set file, these barcodes will be replaced in the design by new random barcodes.

Demultiplexer

This program demultiplexes fastq or fastq.gz files obtained from sequencing with inline barcodes.
Like used in GBS, RAD, ... protocols.

These parameters are mandatory:

• -f1 the name and path of the fastq or fastq.gz file to demultiplex
• -i the name and path of the info file. This is a tab delimeted file without headings, with three (or more) columns: sample, sequence of the barcode, name of the enzyme, name of the second enzyme (optional, can be an empty string), the second barcode (optional, can be an empty string),mismatches for the barcode (optional)

These parameters are optional:

• -f2 the name of the second fastq or fastq.gz file (only with paired-end sequencing)

• -o the name of the output directory (standard the directory of the call)

• -t the number of threads to use (standard 1)

• -lf use long file names (standard false) filename is standard the sample name, long file names is sample name _ barcode _ enzyme

• -rad if the data is rad data or not (-rad true for RAD data, -rad false for GBS data) standard false (GBS)

• -gzip the input and output are/must be gziped (.gz) (standard false: input and output are .fastq, if true this is .fastq.gz)

• -mb the allowed mismatches in the barcodes (overrides the option -m)

• -me the allowed mismatches in the enzymes (overrides the option -m)

• -minsl the minimum allowed length for the sequences (standard 0, rejected sequences are found in the stats for each sample in the rejected.count column. The sequences are found untrimmed in the undetermined file.)

• -n keep sequences where N occurs as a "nucleotide" (standard true)

• -ca the common adaptor used in the sequencing (standard (only first piece) AGATCGGAAGAGCG) currently only used for adaptor ligase see -al and when -rad is true) (minimum length is 10)

• -s the posible distance of the start. This is the distance count from the start of the read to the first basepair of the barcode or enzyme (standard 0, maximum 20)

• -kc Keep the enzyme cut-site remains (standard true) (example: enzyme ApeKI and restriction site G^CWGC: "ApeKI \tab CAGC,CTGC")

• -ea Add enzymes from the given file (keeps the standard enzymes, and add the new) (enzyme file: no header, enzyme name tab cutsites (multiple cutsites are comma separeted)) (only use once, not use -er) (example: enzyme ApeKI and restriction site G^CWGC: "ApeKI \tab CAGC,CTGC")

• -er Replace enzymes from the given file (do not keep the standard enzymes) (enzyme file: no header, enzyme name tab cutsites (multiple cutsites are comma separeted)) (only use once, not use -ea)

• -scb Use self correcting barcodes (barcodes created by the barcodeGenerator) (standard false)

• -malg the used algorithm to find mismatches and indels, possible algorithms:
  * hammings (Standard) Checks for mismatches (no indels)
  * knuth Faster than hammings, but can miss some locations
  * indelmis Checks for mismatches and indels, the barcode/enzyme/ adaptor with the least errors (mismatches or indels) is taken
  * misindel Checks for mismatches and indels, the mismatches are supperior to the indels (faster than indelmis, but errors can be higher)

• -q the kind of quality scores used in the fastq file (including how phred scores are encoded):
  * Illumina1.8 (Standard)
  * Illumina1.5
  * Illumina1.3
  * Sanger
  * Solid

Possible Standard Enzymes for the info file: (NA is no enzyme)

• ApeKI
• PstI
• EcoT22I
• PasI
• HpaII
• MspI
• PstI-EcoT22I
• PstI-MspI
• PstI-TaqI
• SbfI-MspI
• AsiSI-MspI
• BssHII-MspI
• FseI-MspI
• SalI-MspI
• ApoI
• BamHI
• MseI
• Sau3AI
• RBSTA
• RBSCG
• NspI
• AvaII
• NAN

Barcode Discovery

This program searches for possible barcodes and barcode enzyme combinations.
Designed for the discovery of sequencing errors, or unused barcodes when a large proportion of the demultiplex is undetermined.

Mandatory parameters:

• -f1 the name of the input file (mandatory)
  Optional parameters:
• -min the minimum length of the barcode (standard 6)
• -max the maximum length of the barcode (standard 10)
• -gzip use gzip files as input and output (standard false)
• -o the output directory (standard the directory of execution)
• -ea Add enzymes from the given file (keeps the standard enzymes, and add the new) (enzyme file: no header, enzyme name tab cutsites (multiple cutsites are comma separeted)) (example: enzyme ApeKI and restriction site G^CWGC: "ApeKI \tab CAGC,CTGC") (only use once, not use -er)
• -er Replace enzymes from the given file (do not keep the standard enzymes) (enzyme file: no header, enzyme name tab cutsites (multiple cutsites are comma separeted)) (example: enzyme ApeKI and restriction site G^CWGC: "ApeKI \tab CAGC,CTGC") (only use once, not use -ea)
• -barmin The minimum occurance of a barcode before it is shown in the results (standard: 200)
• -barmax The maximum occurance of barcodes shown in the output (increasing this number will increase ram usage, but gives a slightly better result) (standard: 100)
• -barmis The percentage of mismatches that may occure between barcodes (integer between 1 and 10) (standard: 10)

Tutorial

See the Tutorial file and the example folder.

Change Logs

v1.0

• The original version

v1.0.1

• While demultiplexing, the number of demultiplexed reads are shown for every 100000 reads

v1.1

• Possible to simulate and demultiplex dual barcode experiments (in paired end modus only)
• Updated barcode recognition for paired end modus in the demultiplexer: when a read can be assigned to multiple samples, the read is considered as unvalid (previous was first sample)

v1.1.1

• Updated output and stats for dual barcode experiments

v1.1.2

• Updated barcode recognition for single read modus in the demultiplexer: when a read can be assigned to multiple samples, the read is considered as unvalid (previous was first sample)

v1.1.3

• On request added the enzyme AvaII

v1.1.4

• Update adaptor ligase finding algorithm
• Removed unneeded, confusing parameters -cc and -al
• Removed unused code

v1.1.5

• Update digest (removed possible input file parsing error)
• Updated Single Read Demultiplexing

v1.2

• Deleted Demultiplexer option -m
• Code Clean-up

v1.3

• Removal of the filelock class, ReentrantLock is now used
• The demultiplexing is now multithreaded: use -t NumberOfThreads