obitools/obitools4
2
0
Fork 0
mirror of https://github.com/metabarcoding/obitools4.git synced 2025-06-29 08:10:45 +00:00
Code Issues Packages Projects Releases Wiki Activity

obimultiplex2 in now obimultiplex

Browse Source 
Former-commit-id: ee6f5e15a3f1729dfc2806d039c842c9c3bdd343
This commit is contained in:
Eric Coissac
2024-06-07 09:04:37 +02:00
parent bc12b1aaa2
commit 3a7661f324
6 changed files with 118 additions and 235 deletions
Download Patch File Download Diff File Expand all files Collapse all files

56
cmd/obitools/obimultiplex2/main.go
View File

@ -1,56 +0,0 @@
package main
import (
"fmt"
"os"
log "github.com/sirupsen/logrus"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiiter"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obioptions"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obiconvert"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obimultiplex2"
)
func main() {
// f, err := os.Create("cpu.pprof")
// if err != nil {
// log.Fatal(err)
// }
// pprof.StartCPUProfile(f)
// defer pprof.StopCPUProfile()
// ftrace, err := os.Create("cpu.trace")
// if err != nil {
// log.Fatal(err)
// }
// trace.Start(ftrace)
// defer trace.Stop()
optionParser := obioptions.GenerateOptionParser(obimultiplex2.OptionSet)
_, args := optionParser(os.Args)
if obimultiplex2.CLIAskConfigTemplate() {
fmt.Print(obimultiplex2.CLIConfigTemplate())
os.Exit(0)
}
if !obimultiplex2.CLIHasNGSFilterFile() {
log.Error("You must provide a tag list file following the NGSFilter format")
os.Exit(1)
}
sequences, err := obiconvert.CLIReadBioSequences(args...)
if err != nil {
log.Errorf("Cannot open file (%v)", err)
os.Exit(1)
}
amplicons, _ := obimultiplex2.IExtractBarcode(sequences)
obiconvert.CLIWriteBioSequences(amplicons, true)
amplicons.Wait()
obiiter.WaitForLastPipe()
}

111
pkg/obitools/obimultiplex2/options.go → doc/book/wolf_data/wolf_diet_ngsfilter.csv Normal file → Executable file
View File

@ -1,105 +1,4 @@
package obimultiplex2
import (
"fmt"
"os"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiformats"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obingslibrary"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obiconvert"
"github.com/DavidGamba/go-getoptions"
log "github.com/sirupsen/logrus"
)
var _NGSFilterFile = ""
var _askTemplate = false
var _UnidentifiedFile = ""
var _AllowedMismatch = -1
var _AllowsIndel = false
var _ConservedError = false
// PCROptionSet defines every options related to a simulated PCR.
//
// The function adds to a CLI every options proposed to the user
// to tune the parametters of the PCR simulation algorithm.
//
// # Parameters
//
// - option : is a pointer to a getoptions.GetOpt instance normaly
// produced by the
func MultiplexOptionSet(options *getoptions.GetOpt) {
options.StringVar(&_NGSFilterFile, "tag-list", _NGSFilterFile,
options.Alias("t"),
options.Description("File name of the NGSFilter file describing PCRs."))
options.BoolVar(&_ConservedError, "keep-errors", _ConservedError,
options.Description("Prints symbol counts."))
options.BoolVar(&_AllowsIndel, "with-indels", _AllowsIndel,
options.Description("Allows for indels during the primers matching."))
options.StringVar(&_UnidentifiedFile, "unidentified", _UnidentifiedFile,
options.Alias("u"),
options.Description("Filename used to store the sequences unassigned to any sample."))
options.IntVar(&_AllowedMismatch, "allowed-mismatches", _AllowedMismatch,
options.Alias("e"),
options.Description("Used to specify the number of errors allowed for matching primers."))
options.BoolVar(&_askTemplate, "template", _askTemplate,
options.Description("Print on the standard output an example of CSV configuration file."),
)
}
func OptionSet(options *getoptions.GetOpt) {
obiconvert.OptionSet(options)
MultiplexOptionSet(options)
}
func CLIAllowedMismatch() int {
return _AllowedMismatch
}
func CLIAllowsIndel() bool {
return _AllowsIndel
}
func CLIUnidentifiedFileName() string {
return _UnidentifiedFile
}
func CLIConservedErrors() bool {
return _UnidentifiedFile != "" || _ConservedError
}
func CLIHasNGSFilterFile() bool {
return _NGSFilterFile != ""
}
func CLINGSFIlter() (*obingslibrary.NGSLibrary, error) {
file, err := os.Open(_NGSFilterFile)
if err != nil {
return nil, fmt.Errorf("open file error: %v", err)
}
log.Infof("Reading NGSFilter file: %s", _NGSFilterFile)
ngsfiler, err := obiformats.ReadNGSFilter(file)
if err != nil {
return nil, fmt.Errorf("NGSfilter reading file error: %v", err)
}
return ngsfiler, nil
}
func CLIAskConfigTemplate() bool {
return _askTemplate
}
func CLIConfigTemplate() string {
return `###
###
### Example of NGSFilter CSV configuration file
###
#
@ -192,11 +91,11 @@ func CLIConfigTemplate() string {
# The sample tag must be unique in the library for a given pair of primers
# + They can be a simple DNA word as here. This means that the same tag is used
# for both primers.
# + It can be two DNA words separated by a colon. For example, aagtag:gaagtag.
# + It can be two DNA words separated by a colon. For example, `aagtag:gaagtag`.
# This means that the first tag is used for the forward primer and the second for the
# reverse primers. "aagtag" is the same as "aagtag:aagtag".
# + In the two word syntax, if a primer forward or reverse is not tagged, its tag
# is replaced by a hyphen '-', for example 'aagtag:-' or '-:aagtag'.
# is replaced by a hyphen `-`, for example `aagtag:-` or `-:aagtag`.
# For a given primer all the tags must have the same length.
# - forward_primer: the forward primer sequence
# - reverse_primer: the reverse primer sequence
@ -205,6 +104,4 @@ experiment,sample,sample_tag,forward_primer,reverse_primer
wolf_diet,13a_F730603,aattaac,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,15a_F730814,gaagtag,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,26a_F040644,gaatatc,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,29a_F260619,gcctcct,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
`
}
wolf_diet,29a_F260619,gcctcct,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
Can't render this file because it has a wrong number of fields in line 4.

2
pkg/obioptions/version.go
View File

@ -7,7 +7,7 @@ import (
// TODO: The version number is extracted from git. This induces that the version
// corresponds to the last commit, and not the one when the file will be
// commited
var _Commit = "a396240"
var _Commit = "748a235"
var _Version = "Release 4.2.0"
// Version returns the version of the obitools package.

11
pkg/obitools/obimultiplex/demultiplex.go
View File

@ -28,19 +28,20 @@ func IExtractBarcode(iterator obiiter.IBioSequence) (obiiter.IBioSequence, error
log.Fatalf("%v", err)
}
worker := obingslibrary.ExtractBarcodeSliceWorker(ngsfilter, opts...)
worker := ngsfilter.ExtractMultiBarcodeSliceWorker(opts...)
newIter := iterator.MakeISliceWorker(worker, false)
out := newIter
if !CLIConservedErrors() {
log.Println("Discards unassigned sequences")
newIter = newIter.Rebatch(obioptions.CLIBatchSize())
log.Infoln("Discards unassigned sequences")
out = out.FilterOn(obiseq.HasAttribute("demultiplex_error").Not(), obioptions.CLIBatchSize())
}
var unidentified obiiter.IBioSequence
if CLIUnidentifiedFileName() != "" {
log.Printf("Unassigned sequences saved in file: %s\n", CLIUnidentifiedFileName())
unidentified, newIter = newIter.DivideOn(obiseq.HasAttribute("demultiplex_error"),
unidentified, out = newIter.DivideOn(obiseq.HasAttribute("demultiplex_error"),
obioptions.CLIBatchSize())
go func() {
@ -56,5 +57,5 @@ func IExtractBarcode(iterator obiiter.IBioSequence) (obiiter.IBioSequence, error
}
log.Printf("Sequence demultiplexing using %d workers\n", obioptions.CLIParallelWorkers())
return newIter, nil
return out, nil
}

112
pkg/obitools/obimultiplex/options.go
View File

@ -15,7 +15,7 @@ import (
var _NGSFilterFile = ""
var _askTemplate = false
var _UnidentifiedFile = ""
var _AllowedMismatch = int(2)
var _AllowedMismatch = -1
var _AllowsIndel = false
var _ConservedError = false
@ -99,10 +99,112 @@ func CLIAskConfigTemplate() bool {
}
func CLIConfigTemplate() string {
return `experiment,sample,sample_tag,forward_primer,reverse_primer
return `###
### Example of NGSFilter CSV configuration file
###
#
# The CSV file can contain comments starting with the # character
# and empty lines.
# At the top of the file a set of lines of three or four columns and having
# the first column containing @param can be used to define parameters
# for the obimultiplex tool. The structure of these lines is :
#
# @param,parameter_name,parameter_value
# @param,parameter_name,parameter_value1,parameter_value2
#
# The following lines describes the PCR multiplexed in the sequencing library.
# The first line describes the columns of the CSV file and the following lines
# describe the PCR multiplexed.
#
# Five columns are expected :
#
# - experiment: the experiment name
# - sample: the sample (pcr) name
# - sample_tag: the tag identifying the sample
# - forward_primer: the forward primer sequence
# - reverse_primer: the reverse primer sequence
#
# Supplementary columns are allowed. Their names and content will be used to
# annotate the sequence corresponding to the sample, as the key=value; located
# after the @ sign did in the original ngsfilter file format.
#
###
### Description of the parameters
###
#
# The forward_spacer and the reverse_spacer allow to specify the number of
# nucleotide separating the 5' end of the forward or reverse primer respectively
# to the 3' end of the tag. The default value is 0.
#
# The param spacer allows for specify this value for both forward and reverse
# simultaneously. The spacer parameter can also, when used wirh two arguments,
# allow to specify the # the spacer value for a specific primer:
#
# @param,spacer,CAGCTGCTATGTCGATGCTGACT,2
#
@param,forward_spacer,0
@param,reverse_spacer,0
#
# A new method for designing indel proof tag is to not use one of the four
# nucleotides in their sequence and to flank the tag with this fourth nucleotide.
# That nucleotide is the tag delimiter. Similarly, to the spacer value,
# three ways to specify the tag delimiter exist:
# - the forward_tag_delimiter and reverse_tag_delimiter
# - the tag_delimiter in its two forms with one and two arguments
#
@param,forward_tag_delimiter,0
@param,reverse_tag_delimiter,0
#
# Three algorithms are available to math a pair of tags with a sample.
# It is specified using the @matching parameter. The three possible
# values are strict, hamming, and indel. The default value is strict.
# As for previous parameters, forward_matching and reverse_matching can
# be used to specify the matching value for each primer. And spacer
# can be used with two arguments to specify the matching value for
# a specific primer.
#
@param,matching,strict
#
# The primer_mismatches parameter allows to specify the number of errors allowed
# when matching the primer. The default value is 2. The same declination of
# the parameters forward_primer_mismatches and reverse_primer_mismatches exist.
#
@param,primer_mismatches,2
#
# The @indel parameter allows to specify if indel are allowed during the matching
# of the primers to the sequence. The default value is false. forward_indel and
# reverse_indel can be used to specify the value for each primer.
#
@param,indels,false
#
###
### Description of the PCR multiplexed
###
#
# Below is an example for the minimal description of the PCRs multiplexed in the
# sequencing library.
#
# The first line is the column names and must exist.
# Five columns are expected :
# - experiment: the experiment name, that allows for grouping samples
# - sample: the sample (pcr) name
# - sample_tag: the tag identifying the sample
# The sample tag must be unique in the library for a given pair of primers
# + They can be a simple DNA word as here. This means that the same tag is used
# for both primers.
# + It can be two DNA words separated by a colon. For example, aagtag:gaagtag.
# This means that the first tag is used for the forward primer and the second for the
# reverse primers. "aagtag" is the same as "aagtag:aagtag".
# + In the two word syntax, if a primer forward or reverse is not tagged, its tag
# is replaced by a hyphen '-', for example 'aagtag:-' or '-:aagtag'.
# For a given primer all the tags must have the same length.
# - forward_primer: the forward primer sequence
# - reverse_primer: the reverse primer sequence
#
experiment,sample,sample_tag,forward_primer,reverse_primer
wolf_diet,13a_F730603,aattaac,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,15a_F730814,gaagtag:gaatatc,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,26a_F040644,gaatatc:-,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,29a_F260619,-:-,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,15a_F730814,gaagtag,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,26a_F040644,gaatatc,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
wolf_diet,29a_F260619,gcctcct,TTAGATACCCCACTATGC,TAGAACAGGCTCCTCTAG
`
}

61
pkg/obitools/obimultiplex2/demultiplex.go
View File

@ -1,61 +0,0 @@
package obimultiplex2
import (
log "github.com/sirupsen/logrus"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiiter"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obingslibrary"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obioptions"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq"
"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obiconvert"
)
func IExtractBarcode(iterator obiiter.IBioSequence) (obiiter.IBioSequence, error) {
opts := make([]obingslibrary.WithOption, 0, 10)
opts = append(opts,
obingslibrary.OptionAllowedMismatches(CLIAllowedMismatch()),
obingslibrary.OptionAllowedIndel(CLIAllowsIndel()),
obingslibrary.OptionUnidentified(CLIUnidentifiedFileName()),
obingslibrary.OptionDiscardErrors(!CLIConservedErrors()),
obingslibrary.OptionParallelWorkers(obioptions.CLIParallelWorkers()),
obingslibrary.OptionBatchSize(obioptions.CLIBatchSize()),
)
ngsfilter, err := CLINGSFIlter()
if err != nil {
log.Fatalf("%v", err)
}
worker := ngsfilter.ExtractMultiBarcodeSliceWorker(opts...)
newIter := iterator.MakeISliceWorker(worker, false)
out := newIter
if !CLIConservedErrors() {
log.Infoln("Discards unassigned sequences")
out = out.FilterOn(obiseq.HasAttribute("demultiplex_error").Not(), obioptions.CLIBatchSize())
}
var unidentified obiiter.IBioSequence
if CLIUnidentifiedFileName() != "" {
log.Printf("Unassigned sequences saved in file: %s\n", CLIUnidentifiedFileName())
unidentified, out = newIter.DivideOn(obiseq.HasAttribute("demultiplex_error"),
obioptions.CLIBatchSize())
go func() {
_, err := obiconvert.CLIWriteBioSequences(unidentified,
true,
CLIUnidentifiedFileName())
if err != nil {
log.Fatalf("%v", err)
}
}()
}
log.Printf("Sequence demultiplexing using %d workers\n", obioptions.CLIParallelWorkers())
return out, nil
}