search

Finn-Lab / SanntiS

SMBGC Annotation using Neural Networks Trained on Interpro Signatures
Apache License 2.0
21 stars 1 forks source link

readme

Docker Repository on Quay

SanntiS

SMBGC Annotation using Neural Networks Trained on Interpro Signatures

Tool for identifying biosynthetic gene clusters (BGCs) in genomic & metagenomic data

How to use SanntiS?

Conda

Requires:

Installation

conda create -n sanntis sanntis
conda activate sanntis

Basic tests

conda activate sanntis
sanntis test/files/BGC0001472.fna
conda deactivate sanntis

Use precomputed InterProScan outputs

SanntiS can be executed using preprocessed InterProScan outputs along with a GenBank (GBK) file specifying the coding sequences (CDSs). This integration facilitates a streamlined analysis pipeline for bioinformatics applications, allowing for enhanced functionality and user flexibility.

conda activate sanntis
sanntis --ip-file test/files/BGC0001472.fna.prodigal.faa.gff3 test/files/BGC0001472.fna.prodigal.faa.gb
conda deactivate sanntis

Docker:

Get InterProsScan data:

The size of download file is ~24G, the final directory is 16G. Be sure to have enough space
bash ./get_ips_slim.sh

Docker ready to use script:

Only works if "data/" and sanntis_container.py are in the same directory
sanntis_container.py --help
sanntis_container.py [OPTIONS] ARGUMENTS

Docker image shell:

docker -it --entrypoint bash -v <path to SanntiS/docker>/data/:/opt/interproscan quay.io/repository/microbiome-informatics/sanntis
sanntis --help
sanntis [OPTIONS] ARGUMENTS

Ouput

GFF3 format file

The fields in this header are as follows:

seqname: SeqID of contig, as in prodigal output.
source: sanntis version.
feature: Feature type name, i.e. CLUSTER, CLUSTER_border, CDS.
start: Start position of feature
end: End position of feature
score: empty
strand: empty
frame: empty
attributes:
  ID: ordinal ID for the cluster, beginning with 1.
  nearest_MiBIG: MiBIG accession of the nearest BGC to the cluster in the MIBIG space, measured in Dice dissimilarity coefficient.
  nearest_MiBIG_class: BGC class of nearest_MiBIG.
  nearest_MiBIG_diceDistance: Dice dissimilarity coefficient between ID and nearest_MiBIG.
  score: Post-processing probability output.
  partial: Indicates if a CLUSTER is at the edge of the contig. First and second digits represent 5' and 3' end, respectively. Same as in prodigal's `partial`. "0" shows the cluster is not at the edge, whereas a "1" indicates is at that edge, (i.e. a partial cluster).

Sample:

##gff-version 3
DS999642    SanntiSv0.9.0   CLUSTER 1   136970  .   .   .   ID=DS999642_sanntis_1;nearest_MiBIG=BGC0001397;nearest_MiBIG_class=NRP Polyketide;nearest_MiBIG_diceDistance=0.561;partial=10

antiSMASH compatibility

SanntiS prioritises seamless integration with various downstream analysis tools, leveraging a GFF3 file output for broad compatibility. In addition, one of the key features in this regard is the ability to generate an output compatible with antiSMASH, a widely used tool in the BGC analysis ecosystem.

--antismash_output Option

SanntiS has an --antismash_output option. This option allows you to create a JSON file formatted according to the specifications of antiSMASH. 

sanntis --antismash_output True test/files/BGC0001472.fna

Executing the command above produces a file named with the suffix antismash.json facilitating its use in antiSMASH for enriched analysis. Specifically, this file can be uploaded to the antiSMASH web server under 'Data input' > 'Upload extra annotations', allowing for an integrated analytical approach that leverages external annotation data.

Cite

If you use SanntiS make sure to cite the publication Expansion of novel biosynthetic gene clusters from diverse environments using SanntiS

Expansion of novel biosynthetic gene clusters from diverse environments using SanntiS
Santiago Sanchez, Joel D. Rogers, Alexander B. Rogers, Maaly Nassar, Johanna McEntyre, Martin Welch, Florian Hollfelder, Robert D. Finn
bioRxiv 2023.05.23.540769; doi: https://doi.org/10.1101/2023.05.23.540769