GWAS For Bacteria User Manual

GWAS%20for%20bacteria%20user%20manual

User Manual:

Open the PDF directly: View PDF PDF.
Page Count: 8

GWAS for BACTERIA
Sarah Earle, Chieh-Hsi Wu and Daniel J. Wilson (2015)
SNP ANALYSIS
Usage example:
Rscript /path/of/SNP_GWAS_MAIN.R -dataFile dataFile.txt
-phylogeny phylogeneticTree.newick
-ref_fa referenceGenome.fasta -ref_gbk referenceGenome.gbk
-prefix outputPrefix –script_dir /gwasSourceCodes/snpGWAS/
-CFML_prefix cfmlOuputPrefix -run_gemma yes –PCA yes -
externalSoftware /path/of/externalSofware.txt
dataFile
This specifies the path of a tab-delimited text file containing the data of the isolates.
The file contains three columns:
(1) The unique ids of the isolates
(2) The paths to the sequence data files of the isolates in fasta format
(3) The binary phenotype data of the isolates
The first row of the file must be the column names, namely id, filePath, and
phenotype. Note that fasta files are compressed to gzip files. In addition these
genomes have been mapped to the same reference genome.
An example of the contents in text file required for dataFile is shown below.
id filePath phenotype
ecol1 /home/data/ecol/ecol1.fasta.gz 0
ecol2 /home/data/ecol/ecol2.fasta.gz 1
phylogeny
This specifies the path to a newick file that contains a phylogeny of all the isolates
concerned in the GWAS studies. If this input is not provided, then a phylogeny is
built from the sequence data provided for dataFile. If the number of isolates is less
than 100, the phylogeny is built using PhyML, otherwise it is built using RAxML.
Although the option of building the phylogenetic tree on the fly is available, it is
strongly recommended that the users should provide the phylogeny input. This is
because it is a good idea to check whether phylogeny has been properly reconstructed.
If the phylogeny were built on the fly, there would not be an opportunity to check the
phylogeny before resuming the rest of the GWAS analysis. The phylogeny is used for
imputing missing values and therefore can influence the results from the GWAS
analysis.
ref_fa
This specifies the path to the fasta file of the reference genome used to map the
isolates concerned in this study.
ref_gbk
This specifies the path to the genbank file of the reference genome used to map the
isolates concerned in this study.
prefix
This specifies the prefix of the output files.
script_dir
This specifies the path to the directory where the other required R scripts are located.
Therefore all those R scripts must be in the same directory. The R source codes are all
in the same folder (snpGWAS), which is inside the src folder. Therefore it would be
easiest not to move the source codes and specify the path to the snpGWAS folder for
script_dir.
CFML_prefix
This specifies the prefix of the output files produced by ClonalFrameML if the output
files of ClonalFrameML have already been produced. The input value of
CFML_prefix must be different to that of prefix. If CFML_prefix is not specified,
then ClonalFrameML will be run and the output files will have the prefix as specified
for prefix. The output files of ClonalFrameML are used for imputation.
run_gemma
This specifies whether or not to additionally run the analysis with linear mixed
models using GEMMA (yes or no). The default value is no.
PCA
This specifies whether or not to additionally run a PCA analysis using Eigensoft. The
default is no.
externalSoftware
This specifies the path to a tab-delimited file containing the name and paths of the
external software packages used in the analysis. The tab-delimited file should contain
two columns with headers name and path. The name column and path column
respectively specifies the names and paths of the external software packages and non-
R scripts used in the SNP analysis. The name column must contain the following:
ClonalFrameML
GEMMA
RAxML
PhyML
EigenSoft
ConvertPhylipToFasta
The path column contains the corresponding path to the software package.
ClonalFrameML, GEMMA, RAxML, PhyML and EigenSoft are the names of the
software packages that must be installed prior to the GWAS analysis.
ConvertPhylipToFasta refers to ConvertPhylipToFasta.class, which is in the same
directory as the R scripts (snpGWAS).
KMER ANALYSIS
Usage example:
Rscript /path/of/KmerAnalysisMain.R -dataFile dataFile.txt
-srcDir /gwasSourceCodes/kmerGWAS/ -prefix outputPrefix
-signif 100000 -genFileFormat BAM –minCov 5 –kmerFileDir
/home/kmers/ -nproc 5 –createDB –refSeqFasta1 refSeqFasta1.txt
–refSeqFasta2 refSeqFasta2.txt –ncbiSummary ncbiSummary.txt
–runLMM TRUE -relateMatrix relatednessMatrix.txt –signifLMM 100000
-externalSoftware /path/of/externalSofware.txt
or
Rscript /path/of/KmerAnalysisMain.R -dataFile dataFile.txt
-srcDir /gwasSourceCodes/kmerGWAS/ -prefix outputPrefix
-signif 100000 -createKmerFiles TRUE –kmerFileDir /home/kmers/
-nproc 5 –createDB –refSeqFasta1 refSeqFasta1.txt –ncbiSummary
ncbiSummary.txt -db2 /path/to/db2 –runLMM TRUE
-relateMatrix relatednessMatrix.txt –signifLMM 100000!
-externalSoftware /path/of/externalSofware.txt
dataFile
This specifies the path to a tab-delimited file containing the data of the isolates. The
file contains three columns:
(1) The unique ids of the isolates
(2) The paths to the sequence data files of the isolates
(3) The binary phenotype data of the isolates
The first row of the file must be the column names, namely id, filePath, and
phenotype. If the user intends to create the kmer files from bam files then the
sequence data files would be BAM or FASTA files. The filePath column would
therefore contain a list of BAM (or FASTA) file paths.
The following is an example of the contents in the data file required when the user
intends to create the kmer files from BAM files.
id filePath phenotype
ecol1 /home/data/ecol/ecol1.bam 0
ecol2 /home/data/ecol/ecol2.bam 1
If the kmer files have already been created, then the filePath contains a list of paths to
gzipped kmer files. The following shows an example of the contents in the data file
required when the user already has the kmer files at hand.
filePath phenotype
/home/data/ecol/ecol1.kmer.txt.gz 0
/home/data/ecol/ecol2.kmer.txt.gz 1
In the phenotype column, 1 and 0 respectively denote the presence and absence of the
trait of interest.
srcDir
This specifies the path of the directory that contains all the other required R scripts
and therefore all those R scripts must be in the same directory. The R source codes are
all in the same folder (kmerGWAS), which is inside the src folder. Therefore it would
be easiest not to move the R source codes and specify the path to the kmerGWAS
folder for srcDir.
prefix
This is the prefix of the output files created from the kmer-based GWAS analysis.
signif
This is an integer input, which specifies the number of top significant kmers to be
annotated. The default value is 10000.
genFileFormat
This requires a String input, which is the file format of the genomic data. The
accepted values are BAM, FASTA and KMER. If the user has specified BAM or
FASTA, then the kmer files will be generated for the association tests. If the kmer
files have already been generated, then the user should specify KMER for this input.
The file format specified here must be consistent with the files specified in the data
file (for input dataFile). E.g. if the user has specified BAM here, the data file should
contain the bam file paths (see dataFile).
minCov
This specifies the minimum depth of the kmers. If the genomic data is based on
assemblies then this should be 1. The default value is 5.
kmerFileDir
This specifies the path of the directory where the kmer files, that are to be created
from bam files, will be located. The default is current working directory.
nproc
This is an integer input, which specifies the number of CPU processors used for
creating kmer files and kmer annotation. The default is 1 processor.
createDB
This requires a Boolean input, which specifies whether or not to create the blast
databases for annotation. The input value must be either TRUE or FALSE. The
default value is TRUE.
Setting createDB to TRUE
With this setting the user can create up to two nucleotide blast databases and convert
the NCBI summary file to the required format for kmer annotation. The NCBI
summary file must be provided (see more details below on ncbiSummary). If the
user would like to create only one blast database, then the input of refSeqFasta1 must
be provided (see more details below on refSeqFasta1). If the user would like to
create two blast databases, then the inputs of refSeqFasta1 and refSeqFasta2 must
be provided (see more details below on refSeqFasta2).
Setting createDB to FALSE
With this setting the user must specify all the inputs for db1, db2 and ncbiAnnot.
refSeqFasta1
This specfies the path of a text file containing a list of paths to all the genome
sequences to be used to create the first database for kmer annotation. The genome
sequences are in fasta file format.
To create this list, go to http://www.ncbi.nlm.nih.gov/guide/howto/dwn-genome/ and
then click on Genomes FTP site. Click on guest when the box pops up, then click on
connect. Go into the folder of your species, then download the nucleotide fasta files
for your species. The refSeqFasta1 is a text file with a single column containing
paths to all of these fasta files.
refSeqFasta2
This specifies the path of the file that is in the same format as the input for
refSeqFasta1. This file is used to create the second blast database and the genome
sequences used should be different to those used to create the first database.
ncbiSummary
This specifies the path to the summary file downloaded from the NCBI database for
the genes of a given species. To create this summary file,
(1) Go to NCBI gene (http://www.ncbi.nlm.nih.gov/gene/).
(2) Put the species of interest in the search bar, e.g. S. aureus
(3) When the search has been completed and matches have been loaded, at the top
right it should say “Send to:”. Click “Send to:”.
(4) Click “File”.
(5) Choose “Summary (text)”.
(6) Click “Create File”.
db1
This is the path to the first blast database used for kmer annotation. This must be
specified if createDB is set to FALSE or refSeqFasta1 is not specified. The database
would have been created using makeblastdb. makeblastdb would have created several
files with the same name having different extensions. For db1, specify the path of the
output files created by makeblastdb but exclude the extensions.
For example, let the following be the paths to the files created by makeblastdb for the
first blast database:
/home/db/example1.nhr
/home/db/example1.nin
/home/db/example1.nsq
/home/db/example1 would then be the input for db1.
db2
This specifies the path to the second blast database used for kmer annotation. This
must be specified if refSeqFasta2 is not specified. The specification of this option
follows db1.
ncbiAnnot
This specifies the path to the NCBI annotation file created from the NCBI summary
file.
runLMM
This requires a Boolean input, which specifies whether or not to run the analysis with
linear mixed model (LMM). The input value must be either TRUE or FALSE. The
default value is true.
relateMatrix
This specifies the path to a text file containing the relatedness matrix created from
GEMMA. The number of rows and columns of the matrix should be the same as the
number of isolates. If runLMM is TRUE, then the relatedness matrix must be
provided.
signifLMM
This requires an integer input that specifies the number of top significant kmers to be
used for the LMM analysis. The default is 100000.
externalSoftware
This specifies the path of a tab delimited file containing two columns with headers
name and path. The name column and path column respectively specifies the names
and paths of the external software packages and non-R scripts used in the analysis.
The name column must contain the following:
parallel
samtools
samToFastq
trimmomatic
trimmomaticPE
shuffleSequencesFastq
dsk
blastn
makeblastdb
GEMMA
The path column contains the corresponding path to the software package. Table 1
presents the software package to which each name refers.
Table 1: External software packages used in the kmer-based GWAS analysis.
name
Software package
parallel
GNU parallel
samtools
SAMtools
samToFastq
Picard command line tool SamToFastq.jar
trimmomatic
Trimmomatic
trimmomaticPE
The Trimmomatic adaptor file combined-PE.fa
shuffleSequencesFastq
shuffleSequences_fastq.pl in the software package velvet
dsk
DSK
blastn
BLAST
makeblastdb
BLAST
GEMMA
GEMMA
All the software packages mentioned in Table 1 must be installed prior to the GWAS
analysis.
PrintOutTopXChisq , sortDsk and gwasKmerPattern refers to compiled scripts not re-
written in R. These are in the same folder as the R source scripts. Table 2 presents the
file names for those non-R scripts.
Table 2: The file name of the non-R scripts.
name
Software package
PrintOutTopXChisq
PrintOutTopXChisq.class
sortDsk
sort_dsk
gwasKmerPattern
gwas_kmer_pattern

Navigation menu