Set of Linux utilities to validate and manipulate fastq files. It also includes a set of programs to preprocess barcodes (namely UMIs, cells and samples), add the barcodes as tags in BAM files and count UMIs.
Programs:
- fastq_info
- fastq_filterpair
- fastq_filter_n
- fastq_num_reads
- fastq_trim_poly_at
- fastq_pre_barcodes
- bam_add_tags
- bam_umi_count
- fastq2bam
- bam2fastq
An easy way to install fastq_utils is via Conda. See the the Conda documentation to get started if you don't already have a working installation.
Installation of fastq_utils is then as simple as:
conda install -c bioconda fastq_utils
This handles software dependencies and provides the pre-built binaries for fastq_utils itself.
Alternatively you can build the package yourself from source.
samtools (version 0.1.19) and zlib (http://zlib.net) version 1.2.11 or latest are required to compile fastq_utils. The install_deps.sh script in the toplevel folder tries to download and compile the dependencies. The bam_annotate.sh script requires samtools (version 1.5 or higher).
Option 1: download the latest source release tarball from https://github.com/nunofonseca/fastq_utils/releases, and then from your download directory type:
tar xzf fastq_utils-x.x.x.tar.gz
cd fastq_utils-x.x.x
Option 2: to use git to download the repository with the entire code history, type:
git clone https://github.com/nunofonseca/fastq_utils.git
cd fastq_utils
To compile fastq_utils, type:
make && make install
If something goes wrong, then remove the whole build subdirectory with make clean and start new with make. The executables will be installed in the bin folder.
Usage: fastq_info [-s -r ] fastq_file1 [fastq_file2|pe]
If the fastq file(s) pass the checks then the program will exit with an exit status of 0 otherwise an error message is printed describing the error and the exit status will be different from 0. Further details about the checks are available in the wiki page.
By using the -r option no checks are made to determine if the read names/identifiers are unique (fastq_info will run faster and use less memory). The -s option can be used when the reads are sorted in the same way in two paired fastq files. This option combined with -r for paired fastq files will make the validation checks less strict but fastq_info will run faster and use a fraction of the memory.
Check a single fastq file
fastq_info file1.fastq.gz
Check a interleaved fastq file
fastq_info file.fastq.gz pe
Check two paired fastq files
fastq_info file_1.fastq.gz file_2.fastq.gz
There is a script available (fastq_validator.sh) that also accepts a BAM file or bzip2 compressed .FASTQ files as input besides of .FASTQ or .FASTQ.gz files.
fastq_validator.sh file_1.fastq.bzip2 file_2.fastq.bzip2
Usage: fastq_filterpair fastq_file1 fastq_file2 out_fastq_file1.fastq.gz out_fastq_file2.fastq.gz out_fastq_sing.fastq.gz
The reads with a mate in fastq_file1 and fastq_file2 are written, respectively, to out_fastq_file1.fastq.gz out_fastq_file2.fastq.gz. Reads without a mate (singleton) are kept in out_fastq_sing.fastq.gz.
Example
fastq_filterpair file_1.fastq.gz file_1.fastq.gz file_matched_1.fastq.gz file_matched_1.fastq.gz file_singletons.fastq.gz
Usage: fastq_filter_n [-n threshold] fastq_file
Outputs a gziped fastq content where reads with more than the maximum number of allowed uncalled bases are not included. Threshold is the maximum percentage (ranging from 0 to 100) of bases with uncalled bases that a read can have. default value is 0, which means that a read with a single uncalled base would be discarded.
Usage: fastq_num_reads fastq_file
fastq_trim_poly_at - trims poly-A stretches at the 3'-end and poly-T at 5'-end of each read, optionally discarding reads with a length below the given threshold.
Usage: fastq_trim_poly_at --file input_fastq_file --outfile output_fastq_file --min_poly_at_len integer --min_len integer
Example:
fastq_trim_poly_at --file my.fastq.gz --outfile my.trimmed.fastq.gz --min_poly_at_len 10 --min_len 20
fastq_pre_barcodes - preprocess the reads to move the barcodes (UMI, Cell, ...) to the respective readname, optionally discarding reads with bases in the barcode regions below a given threshold.
Run fastq_pre_barcodes --help to get the full list of options.
Example:
fastq_pre_barcodes --read1 my.umi.fastq.gz --outfile1 tmp.fastq.gz --phred_encoding 33 --read1_offset 22 --read1_size -1 --umi_read read1 --umi_size=8 --umi_offset 12
In the above command, the UMIs (starting in the base 12 and with a length of 8 bases) are extracted from the sequences and inserted in the respective read name. The read sequences in the output file includes the bases starting in position 22 until the end of the sequence. The modified readname will have the following format
@STAGS_CELL=[cell]_UMI=[umi]_SAMPLE=[sample]_ETAGS_[ORIGINAL READ NAME]
where [cell], [umi], and [sample] will have the value of the barcode (if available) and [ORIGINAL_READ_NAME] is, as the name suggest, the read name found in the input fastq file.
Given a bam file generated from fastq files preprocessed by fastq_pre_barcodes, bam_add_tags will add UM (UMI), CR (Cell), and BC (sample) tags to each alignment in the BAM file based on the information found in the respective readnames.
Usage: bam_add_tags --inbam input.bam --outbam output.bam [--tx --tx2gx gene2trans_map.tsv --help]
Given a BAM file annotates the alignments with the features defined in the bed files (genes->GX tag, transcritps->TX tag). Alignments can also be annotated as exonic or intronic if they overlap the regions defined in the exon or intron bed file respectively.
Usage: bam_annotate.sh -b bam_file [ -e exon_bed -i intron_bed -g gene_bed -t transcript_bed ]
Given a BAM file with the UM, CR, and BC tags (as produced by bam_add_tags) together with some extra tag. By default the bam_umi_count will count unique UMIs associated to uniquely mapped reads overlapping annotated genes. The GX tag is expected to contain the gene id. If an alignment overlaps multiple features then the UMI count will be partially (1/y) assigned to each feature. The output file (--ucounts) will contain two or more columns (tab-separated): the feature id (gene id by default); cell (if found in the BAM); sample (if found in the bam); and the respective number of unique UMIs (with at least x number of reads, where x is passed in the parameter --min_reads). Alternatively, a Matrix Market file (mtx) file is generated if --ucounts_MM option is passed. A white list of known UMIs can provided using the --known_umi option and a white list of cells provided with the --known_cells option. This is a simpler and faster approach to count UMIs - as an alternative you may want to consider the umis count command available in the umis package which will try to correct the barcodes.
Usage: bam_umi_count --bam in.bam --ucounts output_filename.tsv [--min_reads 0] [--uniq_mapped|--multi_mapped] [--dump file.tsv] [--tag GX|TX] [--known_umi file_one_umi_per_line] [--known_cells file_one_cell_per_line] [--ucounts_MM]
Usage: fastq2bam -s schema -b out_bam -1 fastqfile1 [ -2 fastqfile2 -3 fastqfile3 -4 fastqfile4]
Currently supported schemas are: drop-seq, 10xV1, 10xV1i, 10xV1a, 10xV2, 10xV3, or none. This will be extended in the near future. To get more information type fastq2bam -s selected_schema
Examples:
fastq2bam -s none -b test_se.bam -1 single_end.fastq
fastq2bam -s none -b test_pe.bam -1 x_1.fastq.gz -2 x_2.fastq.gz
fastq2bam -s 10xV1 -b test_10xv1.bam -1 x_1.fastq.gz -2 x_2.fastq.gz
fastq2bam -s 10xV1 -b test_10xv1.bam -1 x_1.fastq.gz -2 x_2.fastq.gz -3 x_I5.fastq.gz
fastq2bam -s 10xV1i -b test_10xv1a.bam -1 x_RA.fastq.gz -2 x_I1.fastq.gz
fastq2bam -s 10xV1a -b test_10xv1a.bam -1 x_RA_1.fastq.gz -2 x_RA_2.fastq.gz -3 x_I1.fastq.gz
fastq2bam -s 10xV1a -b test_10xv1a.bam -1 x_RA_1.fastq.gz -2 x_RA_2.fastq.gz -3 x_I1.fastq.gz -4 x_I2.fastq.gz
fastq2bam -s 10xV2 -b test_10xv2.bam -1 x_R1.fastq.gz -2 x_R2.fastq.gz -3 x_I5.fastq.gz
fastq2bam -s 10xV2 -b test_10xv2.bam -1 x_R1.fastq.gz -2 x_R2.fastq.gz -3 x_I5.fastq.gz -4 x_I1.fastq.gz
fastq2bam -s drop-seq -b test_ds.bam -1 x_1.fastq.gz -2 x_2.fastq.gz
Usage: bam2fastq --bam in.bam --out fastq_prefix [--10xV2 --10xV3]
Converts a bam file generated by fastq2bam into fastq format - the following fastq files may be generated depending on the content of the BAM file: fastq_prefix_1.fastq.gz, fastq_prefix_2.fastq.gz, fastq_prefix_cell.fastq.gz, fastq_prefix_umi.fastq.gz, fastq_prefix_sample.fastq.gz.
If --10xV2 or --10xV3 is passed to bam2fastq then the following files will be generated: fastq_prefix_R1.fastq.gz, fastq_prefix_R2.fastq.gz, and, depending on contents of the BAM file, fastq_prefix_I1.fastq.gz.