Component: AdaptorRemoval

AdaptorRemoval

Removes tags from the reads in fastq/fasta files. Specific tags for each end of the read can be given as input in tags file or they can be predicted. Overrepresented sequences found by FastQC (SeqQC component) can be given as input using QCParser overrepresented output. When using -predict the predicted tags are in the file tags.csv in the output folder, when using -stats the information will be saved in stats.csv.

Version	2.1
Bundle	sequencing
Categories
Authors	Alejandra Cervera (alejandra.cervera@helsinki.fi)
Issue tracker	View/Report issues
Source files	component.xml AdaptorRemoval.sh
Usage	Example with default values AdaptorRemoval( reads = BinaryFile, //tags = BinaryFile, //overSeqs = BinaryFile, extra = "", fastq = true, matrix = "exact", mm3 = 1, mm5 = 1, out_format = 0, percentage = 15, predict = false, stats = false, tag3 = true, tag5 = true, trim = true, verbose = false )

Version

2.1

Bundle

sequencing

Name	Type	Mandatory	Description
reads	BinaryFile	Mandatory	Input files in FASTQ/FASTA format.
tags	BinaryFile	Optional	A file with the tags to be trimmed. The file needs to have tag3 or tag5 as the first field and the tag as the second field.
overSeqs	BinaryFile	Optional	Overrepresented sequences directly obtained from SeqQC component (via QCParser) or file with adaptors or tags that should be trimmed without specifying from which end. Illumina adaptors will be trimmed regardless of the percentage, other overrepresented sequences will only be trimmed if they exceed the percentage parameter.

tags

BinaryFile

Optional

A file with the tags to be trimmed. The file needs to have tag3 or tag5 as the first field and the tag as the second field.

overSeqs

BinaryFile

Optional

Overrepresented sequences directly obtained from SeqQC component (via QCParser) or file with adaptors or tags that should be trimmed without specifying from which end. Illumina adaptors will be trimmed regardless of the percentage, other overrepresented sequences will only be trimmed if they exceed the percentage parameter.

Outputs

Name	Type	Description
arrayOut	Array<BinaryFile>	The key "read" points to the trimmed file if trimming was performed or to the unmodified input read file if no trimming occurred, the "stats" and "tags" will point to the corresponding files, they may or may not exist depending on the parameters chosen.
removalLog	BinaryFile	Log file
predictLog	BinaryFile	Log file

Name

Type

Description

arrayOut

Array<BinaryFile>

The key "read" points to the trimmed file if trimming was performed or to the unmodified input read file if no trimming occurred, the "stats" and "tags" will point to the corresponding files, they may or may not exist depending on the parameters chosen.

removalLog

BinaryFile

Log file

predictLog

BinaryFile

Log file

Parameters

Name	Type	Default	Description
extra	string	""	Give one or more options as a string, e.g. `extra="-trim_within"`. More options on the Tag Cleaner documentation site, or use extra=-help to see the options With this option you can use all of the parameters from the TagCleaner script, but right now it can only be used when trimming, i.e you cannot use it during the prediction step.
fastq	boolean	true	Defines if the input file is FASTQ (true) or FASTA (false).
matrix	string	"exact"	union, subset, or exact depending on how to match the tags to the sequences
mm3	int	1	Maximum number of allowed mismatches at the 3'-end. The independent definition for the 5'- and 3'-end of the reads accounts for the differences in tag sequences due to the limitations of the sequencing method used to generate the datasets. The 3'-end will in most cases show a lower number of matching tag sequences with low number of mismatches due to incomplete or missing tags at the ends of incompletely sequenced fragments. [default: 0]
mm5	int	1	Maximum number of allowed mismatches at the 5'-end. [default: 0]
out_format	int	0	To change the output format, use one of the following options. If not defined, the output format will be the same as the input format. 1 (FASTA only), 2 (FASTA and QUAL) or 3 (FASTQ)
percentage	int	15	The tags have to be overrepresented by this percentage to be trimmed.
predict	boolean	false	Use this option to have TagCleaner predict the tag sequences. It will attempt to predict the tag at either or both sites, if possible. The algorithm implemented for the tag prediction assumes the randomness of a typical metagenome. Datasets that do not contain random sequences from organisms in an environment, but rather contain, for example, 16S data may cause incorrect detection of the tag sequences. However, the tag sequences will most likely be over-predicted and can be redefined by the user prior to data processing. The tag sequence prediction uses filtered base frequencies instead of raw base frequencies. This allows a more accurate prediction as it accounts for incomplete and shifted tag sequences. The output values are separated by tabs with the header line: "#Param Tag_Sequence Tag_Length Percent_Explained". If no tags are reported, then no tags could be identified in the data set. Cannot be used in combination with -tag3 or -tag5 or -stats. When using this option, no trimming will be performed.
stats	boolean	false	Prints the number of tag sequences matching for different numbers of mismatches. In combination with -split, the number of sequences with fragment-to-fragment concatenations is printed as well. The output values are separated by tabs with the header line: "#Param Mismatches_or_Splits Number_of_Sequences Percentage Percentage_Sum". Cannot be used in combination with -predict and require -tag5 or -tag3. If predict is true, it will predict first and then obtains the stats; it cannot be used in combination with trim, so if trim is also true it will only run stats.
tag3	boolean	true	Set to true if tag sequence at 3'-end will be trimmed.
tag5	boolean	true	Set to true if tag sequence at 5'-end will be trimmed.
trim	boolean	true	If true then it trims the tags in the read_tags file.
verbose	boolean	false	Prints status and info messages during processing.

Name

Type

Default

Description

extra

string

Give one or more options as a string, e.g. extra="-trim_within". More options on the Tag Cleaner documentation site, or use extra=-help to see the options With this option you can use all of the parameters from the TagCleaner script, but right now it can only be used when trimming, i.e you cannot use it during the prediction step.

fastq

boolean

true

Defines if the input file is FASTQ (true) or FASTA (false).

matrix

string

"exact"

union, subset, or exact depending on how to match the tags to the sequences

mm3

int

Maximum number of allowed mismatches at the 3'-end. The independent definition for the 5'- and 3'-end of the reads accounts for the differences in tag sequences due to the limitations of the sequencing method used to generate the datasets. The 3'-end will in most cases show a lower number of matching tag sequences with low number of mismatches due to incomplete or missing tags at the ends of incompletely sequenced fragments. [default: 0]

mm5

int

Maximum number of allowed mismatches at the 5'-end. [default: 0]

out_format

int

To change the output format, use one of the following options. If not defined, the output format will be the same as the input format. 1 (FASTA only), 2 (FASTA and QUAL) or 3 (FASTQ)

percentage

int

The tags have to be overrepresented by this percentage to be trimmed.

predict

boolean

false

Use this option to have TagCleaner predict the tag sequences. It will attempt to predict the tag at either or both sites, if possible. The algorithm implemented for the tag prediction assumes the randomness of a typical metagenome. Datasets that do not contain random sequences from organisms in an environment, but rather contain, for example, 16S data may cause incorrect detection of the tag sequences. However, the tag sequences will most likely be over-predicted and can be redefined by the user prior to data processing. The tag sequence prediction uses filtered base frequencies instead of raw base frequencies. This allows a more accurate prediction as it accounts for incomplete and shifted tag sequences. The output values are separated by tabs with the header line: "#Param Tag_Sequence Tag_Length Percent_Explained". If no tags are reported, then no tags could be identified in the data set. Cannot be used in combination with -tag3 or -tag5 or -stats. When using this option, no trimming will be performed.

stats

boolean

false

Prints the number of tag sequences matching for different numbers of mismatches. In combination with -split, the number of sequences with fragment-to-fragment concatenations is printed as well. The output values are separated by tabs with the header line: "#Param Mismatches_or_Splits Number_of_Sequences Percentage Percentage_Sum". Cannot be used in combination with -predict and require -tag5 or -tag3. If predict is true, it will predict first and then obtains the stats; it cannot be used in combination with trim, so if trim is also true it will only run stats.

tag3

boolean

true

Set to true if tag sequence at 3'-end will be trimmed.

tag5

boolean

true

Set to true if tag sequence at 5'-end will be trimmed.

trim

boolean

true

If true then it trims the tags in the read_tags file.

verbose

boolean

false

Prints status and info messages during processing.

Test cases

Test case	Parameters▼	IN reads	IN tags	IN overSeqs	OUT arrayOut	OUT removalLog	OUT predictLog
case1	properties	reads	(missing)	(missing)	arrayOut	(missing)	(missing)
	# Testing AdaptorRemoval component, fastq = true, predict = true, trim = true, verbose = false, mm3 = 1, mm5 = 1
case2	properties	reads	(missing)	(missing)	(missing)	(missing)	(missing)
	# Testing AdaptorRemoval component, fastq = true, predict = true, trim = false, verbose = true, stats = true, tag3 = true, tag5 = true
case3	properties	reads	tags	(missing)	arrayOut	(missing)	(missing)
	# Testing AdaptorRemoval component, fastq = true, predict = false, trim = true, verbose = false, percentage = 25, stats = false, tag5 = true, tag3 = true

Test case

Parameters▼

IN
reads

IN
tags

IN
overSeqs

OUT
arrayOut

OUT
removalLog

OUT
predictLog

case1

properties

reads

(missing)

arrayOut

(missing)

# Testing AdaptorRemoval component,
fastq = true,
predict = true,
trim = true,
verbose = false,
mm3 = 1,
mm5 = 1

case2

properties

reads

(missing)

# Testing AdaptorRemoval component,
fastq = true,
predict = true,
trim = false,
verbose = true,
stats = true,
tag3 = true,
tag5 = true

case3

properties

reads

AdaptorRemoval

Inputs

Outputs

Parameters

Test cases