Oatk is designed for de novo assembly of complex plant organelle genomes using PacBio HiFi data. It can also be used to assemble other simple organelle genomes such as animal mitochondria. The toolkit consists of four major tools. syncasm is a de novo HiFi read assembler using a sparse de Bruijn graph constructed from closed syncmers (Edgar, R. 2021). hmmannot is a HMMER wrapper for convenient annotation of organelle sequences using a pre-built HMM profile database which is available at OatkDB. pathfinder is a tool used for parsing and circularising organelle genomes from the assembled sequences combining the HMM annotations and assembly graph structure. oatk is a wrapper for running syncasm, hmmannot and pathfinder collectively. There is also an auxiliary tool path_to_fasta used to extract FASTA sequences from a GFA format file given a path.
To compile Oatk from source code, you need to have a C compiler, GNU make and zlib development files installed. Download the source code from this repo or with git clone https://github.com/c-zhou/oatk.git. Then type make in the source code directory to compile.
Oatk can also be installed from BioConda via conda install -c bioconda oatk or as a Nextflow nf-core module via nf-core modules install oatk.
- You need the nhmmscan executable to run
oatkandhmmannot. It could be made available to the program through the environmental path or specified as a program parameter. - You also need the organelle gene profile database to run
oatkandhmmannot. It could be downloaded from OatkDB. There are many pre-built databases ready to use. You can also build your own database following the instructions on that page. For land plant organelles, while there are databases for several lower-rank taxonomies, the embryophyta is the recommended one to start with.
The easiest way is to use the oatk wrapper. Internally, it consecutively runs three modules: HiFi read assembly with syncasm, HMM annotation with hmmannot and organelle genome extraction with pathfinder. You can also run three components separately as detailed below.
The oatk program also allows you to extract organelle genomes from your own genome assembly graph built from such as MBG, Spades or Hifiasm. To do this, you need to include the -G option amd the input sequence file(s) needs to be replaced by a GFA file. It should be noted that oatk (and pathfinder) highly relies on the sequence and arc coverage (especially the sequence coverage) to solve the genome structure, so the overlap/string graphs constructed from many long-read assemblers (e.g., Hifiasm) are not ideal for complete organelle genome construction. Also, different assemblers use different tags for sequence and arc coverage in the GFA file. By default, oatk (and pathfinder) uses KC:i, SC:i and EC:i as the tag for kmer, sequence and arc coverage respectively, where the value of kmer coverage is approximately equal to the product of the sequence coverage and the sequence length. This is the default output format for syncasm. When you use GFA inputs from other assemblers, you may need to explicitly pass this information to oatk (and pathfinder) with --kmer-c-tag (or --seq-c-tag) and --edge-c-tag options. For example, with MBG output, you need to specify --kmer-c-tag FC:f and --edge-c-tag ec:i, otherwise, all coverage values will be considered as 1, and will probably lead to incorrect genome structures.
Below are examples of running these programs using the Arabidopsis thaliana organelle genome data download from this Zenodo data repository.
Here is an example to run oatk,
oatk -k 1001 -c 30 -t 8 --nhmmscan /bin/nhmmscan -m embryophyta_mito.fam -p embryophyta_pltd.fam -o ddAraThal4 ddAraThal4_organelle.hifi.fa.gz
Seven optional and one positional parameters are set in this example:
-k specifies the syncmer size (same as the default value).
-c specifies the synmcer coverage threshold. Syncmers with coverages below the threshold will be excluded from the assembly graph. This is a important parameter for organelle genome assembly. A proper choice of this threshold will help filter out the nuclear genome and numts. As an empirical instruction, this parameter can be set as 5-10 times the value of nuclear sequence coverage.
-t specifies the thread number.
--nhmmscan specifies the path to the nhmmscan executable. If not specified, the program will assume it is in the environmental path.
-m specifies the mitochondrion (MT) gene profile database. With this parameter, the program will attempt to parse the mitochondrion genome.
-p specifies the chloroplast (PT) gene profile database. With this parameter, the program will attempt to parse the chloroplast genome.
-o specifies the prefix of the output files.
The positional parameter specifies the input PacBio HiFi data file. The program recognises both FASTA and FASTQ format, plain and gzipped. Multiple input files are allowed.
Upon a successful run, the command will generate these major files:
ddAraThal4.utg.final.gfa the GFA file for the final genome assembly | syncasm
ddAraThal4.annot_mito.txt the MT gene annotation file for assembled sequences | hmmannot
ddAraThal4.annot_pltd.txt the PT gene annotation file for assembled sequences | hmmannot
ddAraThal4.mito.gfa the subgraph for the MT genome | pathfinder
ddAraThal4.mito.bed the gene annotation for the MT sequences | pathfinder
ddAraThal4.mito.ctg.fasta the structure-solved MT contigs | pathfinder
ddAraThal4.mito.ctg.bed the genome annotation for MT contigs | pathfinder
ddAraThal4.pltd.gfa the subgraph for the PT genome | pathfinder
ddAraThal4.pltd.bed the gene annotation for the PT sequences | pathfinder
ddAraThal4.pltd.ctg.fasta the structure-solved PT contigs | pathfinder
ddAraThal4.pltd.ctg.bed the genome annotation for PT contigs | pathfinder
These three programs share many parameters with the oatk wrapper. Unless specified, the parameters used this section is the same as those in the previous section.
Here is an example to run syncasm,
syncasm -k 1001 -c 30 -t 8 -o ddAraThal4 ddAraThal4_organelle.hifi.fa.gz
The major file generated is ddAraThal4.utg.final.gfa for the GFA file of the final genome assembly.
Here is an example to run hmmannot,
hmmannot -t 8 --nhmmscan /bin/nhmmscan -o ddAraThal4.annot_mito.txt embryophyta_mito.fam ddAraThal4.utg.final.gfa
hmmannot -t 8 --nhmmscan /bin/nhmmscan -o ddAraThal4.annot_pltd.txt embryophyta_pltd.fam ddAraThal4.utg.final.gfa
Three optional and two positional parameters are set in this example:
The -t and --nhmmscan options are the same as the oatk. The -o option set the output file name.
The first positional parameter specifies the HMM profile database. Proper database file for the target species should be selected. In this example, the databases for embryophyta were used. Here, 'mito' for mitochondrion and 'pltd' for chloroplast.
The second positional parameter specifies the file of sequences for annotation. Here we use the GFA file generated by syncasm. The program also recognises FASTA and FASTQ format, both plain and gzipped.
Here is an example to run pathfinder,
pathfinder -m ddAraThal4.annot_mito.txt -p ddAraThal4.annot_pltd.txt -o ddAraThal4 ddAraThal4.utg.final.gfa
Three optional and one positional parameters are set in this example:
-m specifies the mitochondrion annotation file. With this parameter, the program will attempt to parse the mitochondrion genome.
-p specifies the chloroplast annotation file. With this parameter, the program will attempt to parse the chloroplast genome.
-o specifies the prefix of the output files.
The positional parameter specifies the assembly graph file.
The major output files are the organelle subgraph files, contig FASTA files and sequence annotation files as listed under the oatk example.
path_to_fasta is a tool used to extract FASTA sequences from a GFA file with a path. For example, path_to_fasta oatk_utg_final.gfa u1+,u2-,u3-,u2+.
Chenxi Zhou, Max Brown, Mark Blaxter, The Darwin Tree of Life Project Consortium, Shane A. McCarthy, Richard Durbin. Oatk: a de novo assembly tool for complex plant organelle genomes. bioRxiv 2024.10.23.619857; doi: https://doi.org/10.1101/2024.10.23.619857