MycoCosm genome downloader

(Note: This project is unfrequently maintained, and is mostly kept here for archival purposes)

Helps to download genomes (fasta + gff) files from MycoCosm.

For detecting biosynthetic regions using fungiSMASH on all downloaded data, see the companion script

Requirements

• ete3 and libxml2. Alternatively, you can use the provided yml file with conda (conda env create --file mycocosmdownloader.yml)
• curl (should already be installed in Mac or Linux systems)
• A JGI account

Usage (tested 2024-02)

Update taxonomy database

As part of the output of this program, a "taxonomy" file is created, which includes the lineage for each species (this lineage is also used to create the folder structure for the output). For this, `ete3`` is used. Use the following command to update the database:

python mycocosm_genome_downloader.py --update

Obtain data files

Obtain the list of genomes, and the list of files:

python mycocosm_genome_downloader.py --getgenomelist --getxml

The list of genomes is a comma-separated value file which enlists all current projects ("portals") in MycoCosm (MycoCosm_Genome_list.csv, Note: it uses ISO-8859-15 encoding); The list of files is stored as MycoCosm_data.xml (Note: this is ~40Mb file which may take several tries to download. After downloading, it is re-formatted to be human-readable).

These files will be stored in the folder specified by the --outputfolder argument (default: 'output'). You will need your credentials to retrieve the file list.

Download data

Once you have the genome and file list, use them to download the data with

python mycocosm_genome_downloader.py --csv [path to MycoCosm_Genome_list.csv] --xml [path to MycoCosm_data.xml]

This will create a folder structure that follows the fungal taxonomy tree as MycoCosm's main page

Optional: use parameter --simulate to only create the structure, but don't download any file.

Note: as of 2022-01, the downloaded files sum about 33 GiB

Output

Each 'leaf' folder within the folder structure represents a genome from MycoCosm and contains the fasta + gff files (compressed as .gz). Additionally, three files are found inside the base output folder:

• JGI_taxonomy.tsv. Columns: Short name (Portal), Accession (same as short name in this case), TaxId (from NCBI), Name (project name), Path (relative path from output folder to Portal), Assembly file, GFF file, lineage (comma-separated)
• JGI_download_list_[date].txt: All the files that were downloaded
• List_gene_gff_filenames.txt: All gff files for each portal

Optional: re-use data

If you already have downloaded data, you can simply use your local copy of the files instead of downloading them. This makes it easier to update with new genomes. First, use option --getprevious, which will scan a base folder with previous results and create a file called previously_downloaded_files.tsv. Use this file with option --previous to skip files already downloaded

Known limitations, notes

• The incorrect gff might have been chosen
• Some old gff files can't be processed correctly by other tools (e.g. bcbio.gff)
• Can't handle dikaryon genomes ("primary/secondary alleles")

More details

Target download folders

These are the folders that are scanned for files, as they appear while in each Portal's download section:

• Assemblies: "Genome Assembly (unmasked)" (previously called "Assembled scaffolds (unmasked)")
• Gene annotations: "Filtered Models ('best')"

Note: Some genomes seem to be only available in the "masked" version (e.g. Pyrtr1, Altbr1, etc.)

Choosing the right annotation file

There are usually a few files to choose from in the "Filtered Models ('best')" folder. The criteria to choose a single file is implemented in the following order:

• In some cases, the choosing was done manually and incorporated to the hardcoded_gff_files.tsv file. If the Portal has an entry here, use the indicated file. This is a work in progress!
• Skip some specific files: Aciri1_meta_GeneCatalog_genes_20111216.gff.gz, Exoaq1_GeneCatalog_20160901.gff3.gz, Exoaq1_GeneCatalog_20160828.gff3.gz, Fonpe1_GeneCatalog_20160901.gff3.gz, Copmic2_FM1_removed_alleles.gff.gz
• Skip files with proteins or secondary_alleles in their name
• Skip files with the following extensions: gtf.gz, tgz and any other that is not gz
• Prefer gff3 files
• If we have two gff3 (or two gff) files, prefer the most recent

Ignored assembly files / portals

• Files with MitoAssembly, MitoScaffolds, PrimaryAssemblyScaffolds or SecondaryAssemblyScaffolds in their name
• Excluded assemblies (hardcoded). Ignore these filenames as they're not related to assemblies, are old versions or are assemblies of meta-samples: 1034997.Tuber_borchii_Tbo3840.standard.main.scaffolds.fasta.gz, Spofi1.draft.mito.scaffolds.fasta.gz, Patat1.draft.mito.scaffolds.fasta.gz, PleosPC9_1_Assembly_scaffolds.fasta.gz, Neuhi1_PlasmidAssemblyScaffolds.fasta.gz, CocheC5_1_assembly_scaffolds.fasta.gz, Alternaria_brassicicola_masked_assembly.fasta.gz, Aciri1_meta_AssemblyScaffolds.fasta.gz, Rhoto_IFO0880_2_AssemblyScaffolds.fasta.gz
• Ignored portals (hardcoded). Metaprojects or old versions: Rhoto_IFO0880_2, Aciri1_meta, Pospl1

Author

https://github.com/jorgecnavarrom