Home

RXN → UniRef90 Gene Family Mapping

This Wiki documents the end-to-end pipeline for creating reaction-to-gene-family (RXN→GF) mapping files from BioCyc, UniProtKB, and UniRef90 databases.

Database Versions

Database	Version	Notes
UniProtKB	2019_01	Swiss-Prot and TrEMBL databases
UniRef90	2019_01	Clustered sequences at 90% identity
BioCyc	v29.0 (December 10, 2024)	Downloaded under necessary license

Prerequisites

• Python 3.x
• Bash shell
• SLURM cluster environment (optional, for parallel processing)
• 100 GB+ disk space

Required Python scripts:

Script	Used in
parse_uniprotkb_dat_v2.py	Step 1
parse_uniref.py	Step 2
map_uniref90_ecpfam_v02.py	Step 2
build_uniref90sql_db.py	Step 2
get_uniref90_members.py	Step 2
uniref_sqlite.py	Step 2, 3
parse_proteins-dat.py	Step 3
parse_enzrxn-dat.py	Step 3
parse_reactions-dat.py	Step 3
join_reactions.py	Step 3
expand_enzrxn.py	Step 3
merge_enzrxn_pfam_rxn.py	Step 3
merge_rxn_ec_pfam.py	Step 3
query_uniref90sql_db.py	Step 4
rxn_bicoyc_merge.py	Step 5
aggregate_rxns/phase1_partition.py	Step 5
aggregate_rxns/phase2_dedup_bucket.py	Step 5
aggregate_rxns/phase3_concat.py	Step 5

Pipeline Overview

UniProtKB ──┐
            ├──► UniRef90 mapping DBs ──┐
UniRef90 ───┘                           │
                                        ├──► Per-species RXN→GF ──► Aggregated RXN→GF files
BioCyc (multi-species) ─────────────────┘

Steps

Step	Page	Description
1	UniProtKB Processing	Extract and parse UniProtKB flat files
2	UniRef90 Processing	Build UniRef90 annotation and member databases
3	BioCyc Processing	Extract per-species reaction and enzyme data
4	Mapping RXNs to Gene Families	SLURM array job mapping each species
5	Aggregating Multi-Species Mappings	Merge per-species results into final files

See Mapping File Usage Guidelines for guidance on which output file to use.