Raptor

A fast and space-efficient pre-filter for querying very large collections of nucleotide sequences

Download and Installation

There may be performance benefits when compiling from source as the build can be optimized for the host system.

Install with bioconda (Linux)

conda install -c bioconda -c conda-forge raptor

Install with brew (Linux, macOS)

brew install brewsci/bio/raptor

Compile from source

git clone https://github.com/seqan/raptor
git submodule update --init

git clone --branch raptor-v1.1.0 --recurse-submodules https://github.com/seqan/raptor

git checkout raptor-v1.1.0

cd raptor
mkdir -p build
cd build
cmake ..
make

export PATH=$(pwd)/bin:$PATH
raptor --version

Example Data and Usage

A toy data set (124 MiB compressed, 983 MiB decompressed) can be found here.

wget https://ftp.imp.fu-berlin.de/pub/seiler/raptor/example_data.tar.gz
tar xfz example_data.tar.gz

After extraction, the example_data will look like:

$ tree -L 2 example_data
example_data
├── 1024
│   ├── bins
│   └── reads
└── 64
    ├── bins
    └── reads

The bins folder contains a FASTA file for each bin and the reads directory contains a FASTQ file for each bin containing reads from the respective bin (with 2 errors). Additionally, mini.fastq (5 reads of all bins), all.fastq (concatenation of all FASTQ files) and all10.fastq (all.fastq repeated 10 times) are provided in the reads folder.

In the following, we will use the 64 data set. To build an index over all bins, we first prepare a file that contains one file path per line (a line corresponds to a bin) and use this file as input:

seq -f "example_data/64/bins/bin_%02g.fasta" 0 1 63 > all_bin_paths.txt
raptor build --kmer 19 --window 23 --size 8m --output raptor.index all_bin_paths.txt

You may be prompted to enable or disable automatic update notifications. For questions, please consult the SeqAn documentation.

Afterwards, we can search for some reads:

raptor search --error 2 --index raptor.index --query example_data/64/reads/mini.fastq --output search.output

The output starts with a header section (lines starting with #). The header maps a number to each input file. After the header section, each line of the output consists of the read ID (in the toy example these are numbers) and the corresponding bins in which they were found:

#0	example_data/64/bins/bin_00.fasta
#1	example_data/64/bins/bin_01.fasta
...
#62	example_data/64/bins/bin_62.fasta
#63	example_data/64/bins/bin_63.fasta
#QUERY_NAME	USER_BINS
0	0
1	0
2	0
3	0
4	0
16384	1
...
1015812	62
1032192	63
1032193	63
1032194	63
1032195	63
1032196	63

For a list of options, see the help pages:

raptor --help
raptor build --help
raptor search --help
raptor upgrade --help

Preprocessing the input

We offer the option to precompute the minimisers of the input files. This is useful to build indices of big datasets (in the range of several TiB) and also allows an estimation of the needed index size since the amount of minimisers is known. Following above example, we would change the build step as follows:

First we precompute the minimisers and store them in a directory:

seq -f "example_data/64/bins/bin_%02g.fasta" 0 1 63 > all_bin_paths.txt
raptor build --kmer 19 --window 23 --compute-minimiser --output precomputed_minimisers all_bin_paths.txt

Then we run the build step again and use the computed minimisers as input:

seq -f "precomputed_minimisers/bin_%02g.minimiser" 0 1 63 > all_minimiser_paths.txt
raptor build --size 8m --output minimiser_raptor.index all_minimiser_paths.txt

The preprocessing applies the same cutoffs as used in Mantis (Pandey et al., 2018). This means that only minimisers that occur more often than the cutoff specifies are included in the output. If you wish to process all minimisers, you can use --disable-cutoffs.

Partitioned indices

To reduce the overall memory consumption, the index can be divided into multiple (a power of two) parts. This can be done by passing --parts n to raptor build, where n is the number of parts you want to create. This will create n files, each representing one part of the index. The --size parameter describes the overall size of the index. For example, --size 8g --parts 4 will create four 2 GiB indices. This will reduce the memory consumption of raptor build and raptor search by approximately 6 GiB, since there will only be one part in memory at any given time. raptor search will automatically detect the parts, and does not need any special parameters.

Upgrading the index (v1.1.0 to v2.0.0)

An old index can be upgraded by running raptor upgrade and providing some information about how the index was constructed.

Authorship and Copyright

Raptor is being developed by Enrico Seiler, but also incorporates much work from other members of SeqAn.

Citation

In your academic works (also comparisons and pipelines) please cite:

• Raptor: A fast and space-efficient pre-filter for querying very large collections of nucleotide sequences; Enrico Seiler, Svenja Mehringer, Mitra Darvish, Etienne Turc, and Knut Reinert; iScience 2021 24 (7): 102782. doi: https://doi.org/10.1016/j.isci.2021.102782

RECOMB 2021

Raptor was presented at the 25th International Conference on Research in Computational Molecular Biology:

• Slides
• Pre-recorded version of the talk

Please see the License section for information on allowed use.

Supplementary

The subdirectory util contains applications and scripts related to the paper.

License

Raptor is open source software. However, certain conditions apply when you (re-)distribute and/or modify Raptor, please see the license.

Sponsorships

Vercel is kind enough to build and host our documentation and even provide preview-builds within our pull requests. Check them out!