nubeam-dedup

nubeam-dedup is a fast and easy-to-use bioinformatics tool removing exact PCR duplicates for sequencing reads, single-end or paired-end. We appreciate your interest in nubeam-dedup. If you use nubeam-dedup, please kindly cite:

Hang Dai and Yongtao Guan, Nubeam-dedup: a fast and RAM-efficient tool to de-duplicate sequencing reads without mapping. Bioinformatics 36(10), P.3254-3256, (2020) DOI: 10.1093/bioinformatics/btaa112

Compiling:

Compile nubeam-dedup

Run the following commands:

On Linux:

foo@bar:~$ wget --no-check-certificate --content-disposition https://github.com/daihang16/nubeamdedup/archive/master.zip
foo@bar:~$ unzip nubeamdedup-master.zip
foo@bar:~$ cd nubeamdedup-master/Linux/

On macOS:

foo@bar:~$ curl -LJO https://github.com/daihang16/nubeamdedup/archive/master.zip
foo@bar:~$ unzip nubeamdedup-master.zip
foo@bar:~$ cd nubeamdedup-master/macOS/

Then:

foo@bar:Linux$ make && make clean
foo@bar:Linux$ ./nubeam-dedup -i1 ../toydata/1.fq.gz -i2 ../toydata/2.fq.gz 1> out.txt 2> log.txt
foo@bar:Linux$ cat log.txt
Output unique read pairs read 1 to nubeamdedup/Linux/1.uniq.fastq
Output unique read pairs read 2 to nubeamdedup/Linux/2.uniq.fastq
foo@bar:Linux$ cat out.txt
69221/142250 read pairs are unique.
foo@bar:Linux$ wc -l *.fastq
276884 1.uniq.fastq
276884 2.uniq.fastq
553768 total

You should see the expected output as above.

We also offer pre-compiled executable file for Linux. The executable file was compiled on Ubuntu 18.04.2 LTS by compiler gcc with the version of 7.4.0 (Ubuntu 7.4.0-1ubuntu1~18.04). C++11 was used.

Usage:

./nubeam-dedup -h gives you the following messages:

./nubeam-dedup [-i -o -d    -i1 -i2 -o1 -o2 -d1 -d2    -s -r -z -h]

Remove exact PCR duplicates for sequencing reads in (gzipped) fastq format.
Produces de-duplicated reads in fastq files.

Parameters for single-end (SE) reads:
--in or -i: Input file name. The parameter is mandatory for SE reads.
--out or -o: Output file name for unique reads. The default is input file name prefix appended with '.uniq.fastq(.gz)', under the current directory.
--duplicate or -d: File name for removed duplicated reads. The parameter is only valid when --remove or -r is set as 1 (see below). The default is input file name prefix appended with '.removed.fastq(.gz)', under the current directory.

Parameters for paired-end (PE) reads:
--in1 or -i1: Input file name for read 1 file. The parameter is mandatory for PE reads.
--in2 or -i2: Input file name for read 2 file. The parameter is mandatory for PE reads.
--out1 or -o1: Output file name for unique read pairs read 1 file. The default is read 1 file name prefix appended with '.uniq.fastq(.gz)', under the current working directory.
--out2 or -o2: Output file name for unique read pairs read 2 file. The default is read 2 file name prefix appended with '.uniq.fastq(.gz)', under the current working directory.
--duplicate1 or -d1: File name for removed duplicated read pairs read 1 file. The parameter is only valid when --remove or -r is set as 1 (see below). The default is input file name prefix appended with '.removed.fastq(.gz)', under the current directory.
--duplicate2 or -d2: File name for removed duplicated read pairs read 2 file. The parameter is only valid when --remove or -r is set as 1 (see below). The default is input file name prefix appended with '.removed.fastq(.gz)', under the current directory.

Miscellaneous parameters:
--strand or -s: Whether take reads from complementary strand into account. Accept boolean 1 (default) or 0.
--remove or -r: Whether output removed duplicated reads. Accept boolean 0 (default) or 1.
--gz or -z: Compression level of output file. Accept integer 0 (default) to 9. If 0 (default), the output data will not be compressed and will be written to plain text file; otherwise, the output data will be written to gzip format file, with the compression level suggested by user. If compression is needed, a compression level less than 3 is recommended as a compromise between speed and compression.

-h: print this help

Examples:

For single-end reads

Consider reads from complementary strand (default)

./nubeam-dedup -i read.fq

The command gives the following output on screen:

Output unique reads to /current/working/directory/read.uniq.fastq
x/y reads are unique.

Do not consider reads from complementary strand
```
./nubeam-dedup -i read.fq -s 0
```

Consider reads from complementary strand (default), output gzipped file with a compression level of 6, output removed duplicated reads

./nubeam-dedup -i read.fq -z 6 -r 1

The command gives the following output on screen:

Output removed duplicated reads to /current/working/directory/read.removed.fastq.gz    
Output unique reads to /current/working/directory/read.uniq.fastq.gz
x/y reads are unique.

For paired-end reads

Consider reads from complementary strand (default)

./nubeam-dedup -i1 read1.fastq.gz -i2 read2.fastq.gz

The command gives the following output on screen:

Output unique read pairs read 1 to /current/working/directory/reads1.uniq.fastq    
Output unique read pairs read 2 to /current/working/directory/reads2.uniq.fastq
x/y read pairs are unique.

Do not consider reads from complementary strand

./nubeam-dedup -i1 read1.fastq.gz -i2 read2.fastq.gz -s 0

Consider reads from complementary strand (default), output gzipped file with a compression level of 2, output removed duplicated reads

./nubeam-dedup -i1 read1.fastq.gz -i2 read2.fastq.gz -z 2 -r 1

The command gives the following output on screen:

Output removed duplicated read pairs read 1 to /current/working/directory/read1.removed.fastq.gz
Output removed duplicated read pairs read 2 to /current/working/directory/read2.removed.fastq.gz    
Output unique read pairs read 1 to /current/working/directory/reads1.uniq.fastq.gz    
Output unique read pairs read 2 to /current/working/directory/reads2.uniq.fastq.gz
x/y read pairs are unique.

Miscellaneous:

A large value (like 6) for -z tag might significantly increase the running time. From Figures 1, 2 and 7 in this post, -z 6 would increase the amount of time by a factor of 2.5-3 compared with -z 1 (with a limited gain regarding to compression ratio); and -z 1 would increase the amount of time by a factor of 2.5 compared with -z 0, which is the default setting of nubeam-dedup. The recommended practice is: either use a smaller compression level (1-3) or do not use the -z tag at all. For the latter choice, if compression was required, pigz could be used after nubeam-dedup finishes---this can significantly accelerate the compression.
For the convenience of users, nubeam-dedup outputs the number of unique reads and total reads to stdout and the output file information to stderr. Use 1> and 2> to redirect the two streams respectively.