PRINSEQ++ is a C++ implementation of the prinseq-lite.pl program. It can be used to filter, reformat or trim genomic and metagenomic sequence data. It is 5X faster than prinseq-lite.pl and uses less RAM thanks to the use of multi-threading and the cboost libraries. It can read and write compressed (gzip) files, drastically reducing the use of hard drive.
You can install from Conda either in its own channel:
conda create -n prinseq-plus-plus -c conda-forge -c bioconda prinseq-plus-plus
conda activate prinseq-plus-plus
prinseq++ -h
or adding it to another channel:
conda install -c conda-forge -c bioconda prinseq-plus-plus
See below to install prinseq++ from source if you don't want to use conda.
-h | -help
Print the help page; ignore other arguments.
-v | -version
Print version; ignore other arguments.
-threads
Nuber of threads to use. Note that if more than one thread is used, output
sequences might not be in the same order as input sequences. (Default=1)
-VERBOSE <int>
Format of the report of filtered reads, VERBOSE=1 prints information only
on the filters that removed sequences. VERBOSE=2 prints numbers for filters
in order (min_len, max_len, min_gc, max_gc, min_qual_score, min_qual_mean,
ns_max_n, noiupac, derep, lc_entropy, lc_dust, trim_tail_left, trim_tail_right,
trim_qual_left, trim_qual_right, trim_left, trim_right) to compare stats of diferent files.
VERBOSE=0 prints nothing.
(Default=1)
***** INPUT OPTIONS *****
-fastq <file>
Input file in FASTQ format. Can also read a compressed (.gz) file.
-fastq2 <file>
Input file in FASTQ format for pair-end reads. Can also read a
compressed (.gz) file.
-FASTA
Input is in fasta format (no quality). Note that the output format is
still fastq by default. Quality will be treated as 31 (A) for all bases.
-phred64
Input quality is in phred64 format. This is for older Illumina/Solexa reads.
***** OUTPUT OPTION *****
-out_format <int>
Set output format. 0 FASTQ, 1 FASTA. (Default=0)
-out_name <string>
For pair-end sequences, the output files are <string>_good_out_R1 and
<string>_good_out_R2 for pairs where both reads pass quality control,
<string>_single_out_R1 and <string>_single_out_R2 for read that passed
quality control but mate did not. <string>_bad_out_R1 and <string>_bad_out_R2
for reads that fail quality controls. [Default = random size 6 string]
-rm_header
Remove the header in the 3rd line of the fastq (+header -> +). This does
not change the header in the 1st line (@header).
-out_gz
Write the output to a compressed file (WARNING this can be really SLOW,
will be fixed in a future release)
-out_good , -out_single , -out_bad,
-out_good2 , -out_single2, -out_bad2
Rename the output files idividually, this overwrites the names given by
-out_name only for the selected files. File extension won't be added
automatically. (TIP: if you don't need a file, set its name to /dev/null)
***** FILTER OPTIONS ******
-min_len <int>
Filter sequence shorter than min_len.
-max_len <int>
Filter sequence longer than max_len.
-min_gc <float>
Filter sequence with GC percent content below min_gc.
-max_gc <float>
Filter sequence with GC percent content above min_gc.
-min_qual_score <int>
Filter sequence with at least one base with quality score below
min_qual_score.
-min_qual_mean <int>
Filter sequence with quality score mean below min_qual_mean.
-ns_max_n <int>
Filter sequence with more than ns_max_n Ns.
-noiupac
Filter sequence with characters other than A, C, G, T or N.
-derep
Filter duplicated sequences. This only remove exact duplicates.
-lc_entropy=[float]
Filter sequences with entropy lower than [float]. [float] should be in
the 0-1 interval. (Default=0.5)
-lc_dust=[float]
Filter sequences with dust_score lower than [float]. [float] should be in
the 0-1 interval. (Default=0.5)
***** TRIM OPTIONS *****
-trim_left <integer>
Trim <integer> bases from the left (5'->3').
-trim_right <integer>
Trim <integer> bases from the right (3'->5').
-trim_tail_left <integer>
Trim poly-A/T tail with a minimum length of <integer> at the
5'-end.
-trim_tail_right <integer>
Trim poly-A/T tail with a minimum length of <integer> at the
3'-end.
-trim_qual_rule <string>
Rule to use to compare quality score to calculated value. Allowed
options are lt (less than), gt (greater than) and et (equal to).
[default: lt]
-trim_qual_left=[float]
Trim recursively from the 3'-end chunks of length -trim_qual_step if the
mean quality of the first -trim_qual_window bases is less than [float].
(Default=20)
-trim_qual_right=[float]
Trim recursively from the 5'-end chunks of length -trim_qual_step if the
mean quality of the last -trim_qual_window bases is less than [float].
(Default=20)
-trim_qual_window [int]
Size of the window used by trim_qual_left and trim_qual_right (Default=5)
-trim_qual_step [int]
Step size used by trim_qual_left and trim_qual_right (Default=2)
-trim_qual_type <string>
Type of quality score calculation to use. Allowed options are min,
mean, max and sum. [default= min]
If you don't want to use conda, there are several other ways that you can install prinseq++. Try one of these and post an issue if you get stuck.
- g++
- make
- boost-devel ( "sudo apt-get install libboost-all-dev" in Ubuntu)
- pthread
If you are just interested in compiling and using PRINSEQ++, download the latest version. You can also download the binary. If you want to edit the source code, clone this repository.
tar -xvf prinseq-plus-plus-1.2.4.tar.gz
cd prinseq-plus-plus-1.2.4
./configure
make
make test
sudo make installgit clone https://github.com/Adrian-Cantu/PRINSEQ-plus-plus.git
cd PRINSEQ-plus-plus
./autogen.sh
./configure
make
make test
sudo make installprinseq++ -h
Prinseq++ was written by Adrian Cantu, with contributions from Jeff Sadural and Katelyn McNair. Rob Edwards helped here and there.