Introduction

dessimozlab/nf-oma-browser-build is a pipeline for building an OMA Browser instance from an OMA (Orthologous MAtrix) analysis. The pipeline converts the output of either a production OMA run or a FastOMA run into the HDF5 files needed to run a omabrowser webserver. The pipeline integrates a lot of additional data, i.e. GO annotations, domain annotations and cross-references to uniprot and refseq.

Pipeline summary

First part of the pipeline is dependend on input, i.e. production / FastOMA. The later steps are common to both input types.

From production OMA pipeline:

1. extract genomes in dataset from Matrix file
2. extract from genome dbs relevant data such as proteins, locus, etc
3. convert Matrix, extract splicing information

From FastOMA:

TODO: implement and write...

Common part

4. convert HOGs, sequences into HDF5 database, build suffix index and kmer-lookup table (in subworkflow IMPORT_HDF5)
5. import domain annotations if available
6. import cross-references from UniProt and RefSeq (subworkflow GENERATE_XREFS)
7. import GO annotations and Ontology

The pipeline produces in the end in the outputDir (default results/) the necessary files to be loaded into a docker-compose managed omabrowser instance.

Parameters

All parameters are described by running --help of the workflow

nextflow run . --help

Parameter	Description	Type	Default
hog_orthoxml	Hierarchcial orthologous groups (HOGs) in orthoxml format	string
matrix_file	OMA Groups file	string
pairwise_orthologs_folder	Pairwise Orthologs (only by Standard OMA pipeline)	string
genomes_dir	Folder containing genomes	string
known_domains	Folder containing known domain assignments files	string
cath_names_path	File containing CATH domain descriptions	string	http://download.cathdb.info/cath/releases/latest-release/cath-classification-data/cath-names.txt
pfam_names_path	File containing Pfam descriptions	string	https://ftp.ebi.ac.uk/pub/databases/Pfam/current_release/Pfam-A.clans.tsv.gz
xref_uniprot_swissprot	UniProtKB/SwissProt annotation in text format	string	https://ftp.ebi.ac.uk/pub/databases/uniprot/knowledgebase/uniprot_sprot.dat.gz
xref_uniprot_trembl	UniProtKB/TrEMBL annotations in text format	string	/dev/null
xref_refseq	Folder containing RefSeq gbff files.	string
go_obo	Gene Ontology OBO file	string	http://purl.obolibrary.org/obo/go/go-basic.obo
go_gaf	Gene Ontology annotations (GAF format)	string	https://ftp.ebi.ac.uk/pub/databases/GO/goa/UNIPROT/goa_uniprot_all.gaf.gz

Corresponding files from OMA Production Run

Parameter	Path / Value from OMA analysis
hog_orthoxml	HOG in orthoxml format from $DARWIN_OMADATA_PATH/HOGs/, usually latest one
matrix_file	OMA Groups Matrix file, same run as HOG, located in $DARWIN_OMADATA_PATH/Matrix; use the _merged version
genomes_dir	corresponds to $DARWIN_GENOMES_PATH. Must contain Summaries.drw, and all the databases in the subfolders`
pairwise_orthologs_folder	Base directory where the .orth.tsv.gz files have been created. Usually corresponds to $DARWIN_OMA_SCRATCH_PATH/Phase4/. If not specified, no VPairs will be imported and dotplot won't work
known_domains	folder with processed known domain annotations. this will likely change in the future. For now, these files have to be generated outside the pipeline