ADAPTERREMOVAL(1)

📦
version 2.3.4 · 3 versions
, in debian trixie, ubuntu questing, ubuntu resolute, ubuntu stonking

✨ NAME ✨

AdapterRemoval(1) - Fast short-read adapter trimming and processing 💖

🎀 SYNOPSIS 🎀

AdapterRemoval [options...] --file1 <filenames> [--file2 <filenames>]

🦋 DESCRIPTION 🦋

okay so AdapterRemoval is literally the bestie for yeeting leftover adapter sequences off your FASTQ reads 🔥. It works on single-end (SE) or paired-end (PE) reads. You can optionally trim off Ns and low-quality bases, and/or collapse overlapping paired-end mates into one read ✨. Low quality reads get filtered based on their length after trimming and how many ambiguous nucleotides (N) they still have 💅.

All of that can be combined with demultiplexing at the same time, using 5' barcode sequences 🧚. Or, if you're feeling extra, AdapterRemoval can try to rebuild a consensus adapter sequence from paired-end data, so you can figure out which adapters were originally used 👑.

If you use this program, please cite the paper 🫶: Schubert, Lindgreen, and Orlando (2016). AdapterRemoval v2: rapid adapter trimming, identification, and read merging. BMC Research Notes, 12;9(1):88

The paper lives at http://bmcresnotes.biomedcentral.com/articles/10.1186/s13104-016-1900-2 📚

For the full tea, check the detailed docs at http://adapterremoval.readthedocs.io/ 🌈

💅 OPTIONS 💅

  • --help — pulls up a summary of all the command-line options 📋
  • --version — prints the version string 🌸
  • --file1 filename [filenames...] — reads FASTQ reads from one or more files, whether they're uncompressed, bzip2 compressed, or gzip compressed. This is either your SE reads or, in PE mode, your mate 1 reads. If you're running PE, both --file1 and --file2 must be set. Check the primary docs for the list of supported formats 🫧
  • --file2 filename [filenames...] — reads one or more FASTQ files with mate 2 reads for a PE run. If you set this, --file1 has to be set too 🦋
  • --identify-adapters — tries to build a consensus adapter sequence from fully overlapping pairs of PE reads. The minimum overlap is controlled by --minalignmentlength. The result gets compared with --adapter1 and --adapter2. No trimming happens in this mode. Default is off 🔍
  • --threads n — max number of threads. Defaults to 1 🧵

🌸 FASTQ options 🌸

  • --qualitybase base — the Phred quality score encoding used in your input reads. Use 64 for Phred+64 (Illumina 1.3+ and 1.5+) or 33 for Phred+33 (Illumina 1.8+). The value solexa works for reads with Solexa encoded scores too. Default is 33 ✨
  • --qualitybase-output base — the quality score base for reads AdapterRemoval writes out. Same options as before: 64 (Phred+64, i.e., Illumina 1.3+ and 1.5+), 33 (Phred+33, Illumina 1.8+), or solexa. Heads up, quality scores are stored as Phred scores internally, so converting to and from Solexa loses information 🙈. The default matches whatever you gave to --qualitybase
  • --qualitymax base — the max Phred score expected in your input files, which also gets used when writing output. Possible values are 0 to 93 for Phred+33 files, and 0 to 62 for Phred+64 files. Defaults to 41 💜
  • --mate-separator separator — the character that separates the mate number (1 or 2) from the read name in FASTQ records. Defaults to / 🍓
  • --interleaved — turns on --interleaved-input and --interleaved-output together 🧚
  • --interleaved-input — expects the input to be an interleaved FASTQ file given with --file1, where pairs of reads sit right next to each other (e.g. read1/1, read1/2, read2/1, read2/2, etc.) 🌈
  • --interleaved-ouput — writes PE reads into one single file, interleaving mate 1 and mate 2 reads. By default this file is named basename.paired.truncated, but you can change that with the --output1 option 🎀
  • --combined-output — writes all reads into the files specified by --output1 and --output2. Reads that got tossed because of quality filters or read merging get replaced with a single N with Phred score 0. You can combine this with --interleaved-output to write PE reads to a single output file set with --output1 👑

🫧 Output file options 🫧

  • --basename filename — the prefix used for naming output files, unless those names have been overridden with the matching command-line option (see below) 🍓
  • --settings file — an output file with info on the parameters used in the run, plus overall stats on the reads after trimming. Default filename is basename.settings 📋
  • --output1 file — an output file with your trimmed mate1 reads. Default filename is basename.pair1.truncated for paired-end reads, basename.truncated for single-end reads, and basename.paired.truncated for interleaved paired-end reads 💖
  • --output2 file — an output file with trimmed mate 2 reads, as long as --interleaved-output isn't enabled. Default filename is basename.pair2.truncated in PE mode 🦋
  • --singleton file — an output file for paired reads whose mate got discarded. Default filename is basename.singleton.truncated 🥀
  • --outputcollapsed file — if --collapsed is set, this holds overlapping mate pairs that got merged into one read (PE mode), or reads where the adapter was spotted by a minimum overlap, which means the whole template molecule is there. It leaves out reads that later got trimmed for low quality or ambiguous nucleotides. Default filename is basename.collapsed 🔥
  • --outputcollapsedtruncated file — collapsed reads (see --outputcollapsed) that got trimmed because of low-quality or ambiguous nucleotides. Default filename is basename.collapsed.truncated ✂️
  • --discarded file — holds reads discarded because of --minlength, --maxlength or --maxns. Default filename is basename.discarded 🗑️

🧊 Output compression options 🧊

  • --gzip — if set, every FASTQ file AdapterRemoval writes gets gzip compressed, using the level from --gzip-level. The .gz extension gets added to files that didn't get a filename on the command line. Defaults to off 🗜️
  • --gzip-level level — the compression level used when gzipping FASTQ files. Must be a value from 0 to 9, where 0 disables compression and 9 is the best compression. Defaults to 6 ⭐
  • --bzip2 — if set, every FASTQ file AdapterRemoval writes gets bzip2 compressed, using the level from --bzip2-level. The .bz2 extension gets added to files that didn't get a filename on the command line. Defaults to off 🌊
  • --bzip2-level level — the compression level used when bzip2-ing FASTQ files. Must be a value from 1 to 9, where 9 is the best compression. Defaults to 9 🫧

💜 FASTQ trimming options 💜

  • --adapter1 adapter — the adapter sequence you expect to find in mate 1 reads, written in read direction. For a full breakdown of how to give the right adapter sequences, see the "Adapters" section of the online docs. Default is AGATCGGAAGAGCACACGTCTGAACTCCAGTCACNNNNNNATCTCGTATGCCGTCTTCTGCTTG 🧚
  • --adapter2 adapter — the adapter sequence you expect to find in mate 2 reads, written in read direction. Same deal, check the "Adapters" section of the online docs for the details. Default is AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTAGATCTCGGTGGTCGCCGTATCATT 🌸
  • --adapter-list filename — reads one or more adapter sequences from a table. The first two columns (separated by whitespace) of each line should match the values you'd pass to --adapter1 and --adapter2. In SE mode, only column one is needed. Lines starting with # get ignored. When there are multiple rows, AdapterRemoval tries each adapter (pair) and picks the best aligning adapters for each FASTQ read it processes 📋
  • --minadapteroverlap length — in SE mode, reads only get trimmed if the overlap between the read and the adapter is at least X bases long, not counting ambiguous nucleotides (N). This is independent of --minalignmentlength when you use --collapse. It lets you be conservative about picking putative complete inserts in SE mode, while still making sure all possible adapter contamination gets trimmed. The default is 0 🫶
  • --mm mismatchrate — the allowed fraction of mismatches in the aligned region. If the value is less than 1, it gets used directly. If --mismatchrate is greater than 1, the rate is set to 1 / --mismatchrate. The default is 3 when trimming adapters, which corresponds to a max mismatch rate of 1/3, and 10 when using --identify-adapters 💅
  • --shift n — to allow for missing bases at the 5' end of the read, the program can let the alignment slip --shift bases at the 5' end. This is the same as starting the alignment up to --shift nucleotides into read2 (for PE) or the adapter (for SE). The default is 2 🦋
  • --trim5p n [n] — trims the 5' end of reads by a fixed amount after removing adapters, but before quality-based trimming. Give one value to trim mate 1 and mate 2 reads the same amount, or two values separated by a space to trim each mate by a different amount. Off by default 🌈
  • --trim3p n [n] — trims the 3' end of reads by a fixed amount. Works the same way as --trim5p 👑
  • --trimns — trims consecutive Ns from the 5' and 3' termini. If quality trimming is also on (--trimqualities), then stretches of mixed low-quality bases and/or Ns get trimmed too 🔥
  • --maxns n — discards reads containing more than --max ambiguous bases ('N') after trimming. Default is 1000 🗑️
  • --trimqualities: okay so this one trims runs of low quality bases off both the 5' and 3' ends 🧼 The cutoff is set by --minquality. If you've also turned on --trimns, then stretches that mix low-quality bases and Ns get trimmed too 💅
  • --trimwindows window_size: this is a sliding window trimmer, inspired by sickle 🪟 Pass it a window_size and it does the job. There's a whole "Window based quality trimming" section below if you want the deets 🦋
  • --minquality minimum: this sets the quality threshold for --trimqualities and --trimwindows 🎯 The default is 2.
  • --preserve5p: bases on the 5p are off-limits for --trimns, --trimqualities, and --trimwindows 🙅‍♀️ Heads up, collapsed reads won't get quality trimmed when this is on 🫶
  • --minlength length: reads shorter than this get tossed after trimming 🗑️ The default is 15.
  • --maxlength length: reads longer than this get tossed after trimming 📏 The default is 4294967295.

🧬 FASTQ merging options

  • --collapse: for paired-end reads, this merges overlapping mates into one read and recalculates the quality scores 👯 For single-end reads, it tries to find templates where the whole sequence is available 🔍 Either way, the finished collapsed reads get an M_ name prefix, and any collapsed reads that get trimmed because of quality settings get an MT_ prefix 🏷️ The overlap has to be at least --minalignmentlength nucleotides, and the max number of mismatches is set by --mm 💖
  • --minalignmentlength length: this is the minimum overlap between mate 1 and mate 2 before they get collapsed into one (paired-end), or the minimum when hunting for complete template sequences in single-end mode 🧩 The default is 11.
  • --seed seed: when collapsing reads, if the two reads disagree at a position but both bases have the same quality, AdapterRemoval picks one at random 🎲 This option sets the seed for that random number generator, and the value also gets written to the settings file 📝 Heads up: setting the seed isn't reliable in multithreaded mode, since the order of operations is non-deterministic 😭
  • --collapse-deterministic: this turns on deterministic mode 🔒 Right now it only affects --collapse. Overlapping bases that differ but have equal quality get set to N with quality 0, instead of being randomly sampled 🙈 Using this also turns on --collapse for you ✨
  • --collapse-conservatively: this is a different merging algorithm, inspired by FASTQ-join 🦋 For overlapping bases that match, the highest quality score wins 🏆 For overlapping bases that mismatch, the higher quality base wins, and the quality becomes the absolute difference in Phred-score between the two bases 🧮 If mismatching bases have identical quality scores, the base becomes N and the quality score becomes 0 (Phred-encoded) 💀 Using this also turns on --collapse 💅

🧺 FASTQ demultiplexing options

  • --barcode-list filename: this does demultiplexing using a table of one or two fixed-length barcodes, for SE or PE reads 🧺 The table should have 2 or 3 columns. The first column is the sample name, and the second and third are the mate 1 barcode and the (optional) mate 2 barcode 🏷️ For the full scoop, check the "Demultiplexing" section of the online documentation 📚
  • --barcode-mm n: this is the max number of mismatches allowed when counting mismatches in both the mate 1 and mate 2 barcodes for paired reads 🔢
  • --barcode-mm-r1 n: this is the max number of mismatches allowed for the mate 1 barcode 🎀 If you don't set it, it matches the --barcode-mm value, and it can't be higher than --barcode-mm 🙅‍♀️
  • --barcode-mm-r2 n: same deal, but for the mate 2 barcode 🎀 If you don't set it, it matches --barcode-mm, and it can't go higher than --barcode-mm either 🙅‍♀️
  • --demultiplex-only: this only does demultiplexing, using the barcode list you gave --barcode-list. No other processing happens 🧚

✨ WINDOW BASED QUALITY TRIMMING ✨

As of v2.2.2, AdapterRemoval has a sliding window approach to quality-based base trimming, inspired by sickle 🪟 If window_size is greater than or equal to 1, that number is the window size for all reads 🎯 If window_size is a number that's at least 0 but less than 1, then it gets multiplied by the length of each individual read to figure out the window size 🧮 If the window length ends up zero, or bigger than the current read length, the read length gets used instead 📏

Here's how reads get trimmed for a given window size 💅

  1. The new 5' is found by locating the first window where both the average quality and the quality of the first base in that window are greater than --minquality 🌅
  2. The new 3' is found by sliding the first window right, until the average quality drops to less than or equal to --minquality. The new 3' goes on the last base in that window where the quality is greater than or equal to --minquality 🌙
  3. If no 5' position can be determined, the read gets discarded 💀

✨ EXIT STATUS ✨

AdapterRemoval exits with status 0 if the program ran successfully 🥳 If any errors came up, it exits with a non-zero code 🚨 Do not use the output from AdapterRemoval if it returned a non-zero exit code! 🙅‍♀️💀

✨ REPORTING BUGS ✨

Found a bug? Report it on the AdapterRemoval issue tracker 🐛

https://github.com/MikkelSchubert/adapterremoval/issues

✨ LICENSE ✨

This program is free software 🗽 You can redistribute it and/or modify it under the terms of the GNU General Public License, as published by the Free Software Foundation, either version 3 of the License or (your call) any later version 🌈

This program is distributed in the hope that it'll be useful, but it comes with absolutely no warranty 🙈 Not even the implied warranties of merchantability or fitness for a particular purpose. Check out the GNU General Public License for the details 📜

You should have received a copy of the GNU General Public License along with this program. If not, see <http://www.gnu.org/licenses/> 🔗

✨ AUTHOR ✨

Mikkel Schubert and Stinus Lindgreen 👑

2017, Mikkel Schubert and Stinus Lindgreen 🎀

Okay, that's the whole thing, bestie 🫶 You're officially a manpage legend now 🧚 AdapterRemoval v2.3.4, last updated Aug 22, 2024 💖

📦 where this page comes from