mircare/brewery repository overview

$ git clone https://github.com/mircare/Brewery/ --depth 1 && rm -rf Brewery/.git

# To exploit HHblits only (for fast and accurate predictions)
$ python3 Brewery/Brewery.py -i Brewery/example/2FLGA.fasta --cpu 4 --fast

# To exploit both PSI-BLAST and HHblits (for very accurate predictions)
$ python3 Brewery/Brewery.py -i Brewery/example/2FLGA.fasta --cpu 4

# To split a FASTA file with multiple sequences (Optionally)
$ python3 Brewery/split_fasta.py many_sequences.fasta

# To predict all the fasta files in a given directory (Fastas)
$ python3 Brewery/multiple_fasta.py -i Fastas/ --cpu 4 --fast

# To run multiple predictions in parallel (using a total of 8 cores)
$ python3 Brewery/multiple_fasta.py -i Fastas/ --cpu 4 --parallel 2 --fast

$ python3 Brewery/Brewery.py --help
usage: Brewery.py [-h] [-input fasta_file] [--cpu CPU] [--fast] [--noSS]
                  [--noTA] [--noSA] [--noCD] [--distill] [--setup]

This is the standalone of Brewery5. Run it on a FASTA file to predict its
Secondary Structure in 3- and 8-classes (Porter5), Solvent Accessibility in 4
classes (PaleAle5), Torsional Angles in 14 classes (Porter+5) and Contact
Density in 4 classes (BrownAle).

optional arguments:
  -h, --help         show this help message and exit
  -input fasta_file  FASTA file containing the protein to predict
  --cpu CPU          How many cores to perform this prediction
  --fast             Use only HHblits (skip PSI-BLAST)
  --bfd              Harness also the BFD database (https://bfd.mmseqs.com/)
  --noSS             Skip Secondary Structure prediction with Porter5
  --noTA             Skip Torsional Angles prediction with Porter+5
  --noSA             Skip Solvent Accessibility prediction with PaleAle5
  --noCD             Skip Contact Density prediction with BrownAle5
  --distill          Generate useful outputs for 3D protein structure prediction
  --setup            Initialize Brewery5 from scratch (it is recommended when
                     there has been any change involving PSI-BLAST, HHblits,
                     Brewery itself, etc).

E.g., run Brewery on 4 cores: python3 Brewery5.py -i example/2FLGA --cpu 4

# Set the absolute PATHs for databases and query sequences (stored locally)
$ docker run --name brewery -v /**PATH_to_uniprot20_2016_02**:/uniprot20 \
-v /**PATH_to_UniRef90_optional**:/uniref90 -v /**PATH_to_fasta_to_predict**:/Brewery/query \
--cap-add IPC_LOCK mircare/brewery sleep infinity &

# Run a prediction using 5 cores and HHblits only
$ docker exec brewery python3 Brewery.py -i query/2FLGA.fasta --cpu 5 --fast

@article{torrisi_brewery_2020,
	title = {Brewery: Deep Learning and deeper profiles for the prediction of 1D protein structure annotations},
	doi = {10.1093/bioinformatics/btaa204},
	journal = {Bioinformatics},
	author = {Torrisi, Mirko and Pollastri, Gianluca}
}