This code is associated with the paper from Guthrie et al., "The human gut chemical landscape predicts microbe-mediated biotransformation of foods and drugs". eLife, 2019. http://dx.doi.org/10.7554/eLife.42866
The gut microbiota complements host-driven metabolism of dietary compounds and xenobiotics, such as therapeutic drugs. Some gastrointestinally-routed drugs with uncharacterized metabolism and downstream effects on patients, have overlapping chemical properties with dietary compounds, and thus may share gut microbiota enzymatic susceptibilities.
Here, we present a database, MicrobeFDT, which predicts food and drug susceptibility to microbial metabolism by clustering chemically similar drug and food compounds and linking these compounds to microbial enzymes and known toxicities. We hypothesize that the chemical structures of dietary compounds metabolized by bacteria can be used to infer microbial metabolism of uncharacterized drugs with overlapping chemical similarity. We identify therapeutic drugs with potential microbiota-mediated toxicity, highlight understudied microbial enzyme families that may influence drug metabolism and identify drug-food-microbiome interactions that may alter microbiome function and influence host phenotypes.
Instructions for installing docker
Download MicrobeFDT database and configuration
mkdir MicrobeFDT_db
cd MicrobeFDT_db
wget https://www.dropbox.com/s/3xqdrhhamjxhja8/microbefdt.tar
tar xvf microbefdt.tar
ls
conf data microbefdt.tar?dl=0Run neo4j docker image with MicrobeFDT database
Replace $fullpathto with the full path to the data/ and conf/ directories.
docker run --publish=7474:7474 --publish=7687:7687 --volume=/*$fullpathto*/MicrobeFDT_db/data/:/data --volume=/$fullpathto/MicrobeFDT_db/conf:/conf neo4j:3.4.4
Active database: graph.db
Directories in use:
home: /var/lib/neo4j
config: /var/lib/neo4j/conf
logs: /var/lib/neo4j/logs
plugins: /var/lib/neo4j/plugins
import: /var/lib/neo4j/import
data: /var/lib/neo4j/data
certificates: /var/lib/neo4j/certificates
run: /var/lib/neo4j/run
Starting Neo4j.
2018-09-28 15:27:52.495+0000 INFO ======== Neo4j 3.4.4 ========
2018-09-28 15:27:52.526+0000 INFO Starting...
2018-09-28 15:27:55.624+0000 INFO Bolt enabled on 0.0.0.0:7687.
2018-09-28 15:27:58.358+0000 INFO Started.
2018-09-28 15:27:59.575+0000 INFO Remote interface available at http://localhost:7474/Go to http://localhost:7474/ in your browser.
Neo4j is an open source graph database. To interact with the MicrobeFDT database you will need to download Neo4j and follow their instructions for your operating system. MicrobeFDT is run on the Community Edition 3.4.4.
The database file is hosted here.
cd neo4j-community-3.4.4/data/databases
wget https://www.dropbox.com/s/0w85jstzubbk7yz/MicrobefdtNeo4jgraph.zip?dl=0 or download this file from the link.
unzip MicrobefdtNeo4jgraph.zip?dl
The guide file is hosted here.
cd neo4j-community-3.4.4/
wget https://www.dropbox.com/s/z3e917ovxx1xwcm/microbeFDT_guide.zip
mkdir guides
cd guides
unzip microbeFDT_guide.zip
http-server.py microbeFDT.html microbeFDT_guide.zip./bin/neo4j console
Starting Neo4j.
2018-07-31 10:51:31.661+0000 INFO ======== Neo4j 3.4.4 ========
2018-07-31 10:51:31.750+0000 INFO Starting...
2018-07-31 10:51:35.761+0000 INFO Bolt enabled on 127.0.0.1:7687.
2018-07-31 10:51:39.992+0000 INFO Started.
2018-07-31 10:51:41.831+0000 WARN Low configured threads: (max={} - required={})={} < warnAt={} for {}
2018-07-31 10:51:41.850+0000 INFO Remote interface available at http://localhost:7474/
By default your login and password credentials are:
login: neo4j
password: neo4j
After logging in you will be prompted to change your password.
For help getting started run the microbefdt guide.
In a new terminal window navigate to the guides directory.
cd neo4j-microbeFDT/guides
Run the local python server.
python http-server.py
In the neo4j browser run:
:play http://localhost:8001/microbeFDT.html
To recreate the database from source or to access flatfiles used to create the database use the following link.
| Node | Properties | Property Description |
|---|---|---|
| Compound | cpID | PubChem CID |
| name | Compound name | |
| type_OtherMultiCpd_drug | Status as therapeutic drug or other | |
| Enzyme | enzymeID | Unique ID code |
| ec_class | Enzyme Class | |
| ecsd_score | Enzyme specific dominance scores | |
| enzymeECnum | Enzyme Commision ID | |
| hmpGenomeCount | Count across Genomes in HMP | |
| hmpMetagenomeCount | Count across Metagenomes in HMP | |
| Use | ATC_mainGroup | 1st level, anatomical main group |
| ATC_TherapeuticGroup | 2nd level, therapeutic subgroup | |
| ATC_Pharmaologicalsub | 3rd level, pharmacological subgroup | |
| ATC_ChemicalGroup | 4th level, chemical subgroup | |
| ATC_codeID | 5th level, chemical substance |
| Relationship | Properties | Property Description |
|---|---|---|
| HAS_SUBSTRUCTURE_OVERLAP_WITH | score | Pairwise substructure overlap score |
| HAS_TOX_OVERLAP_WITH | score | Pairwise toxicity overlap score |
| METABOLIZED_BY | ||
| TREATMENT_FOR |
MATCH p=()-[r:METABOLIZED_BY]->(n:Enzyme)
WHERE n.enzymeECnum = '3.2.1.31'
RETURN p LIMIT 100
MATCH p=(c:Compound)-[r:HAS_TOX_OVERLAP_WITH]->(c1:Compound)
WHERE r.score > 0.5
RETURN c.cpdID,c1.cpdID,r.score
There are a number of driver that provide access to Neo4j from a variety of supported languages.
Running this example assumes that neo4jis running in the background.
library(RNeo4j)graph = startGraph("http://127.0.0.1:7474/db/data/", username = "neo4j", password = "microbetest")
graphquery = "MATCH p=()-[r:METABOLIZED_BY]->(n:Enzyme) RETURN n.enzymeID,n.ec_class,n.ecsd_score"
cypher(graph, query) n.enzymeID n.ec_class n.ecsd_score
1 (23-dihydroxybenzoyl)adenylate-synthase Transferase 0.6515893
2 (23-dihydroxybenzoyl)adenylate-synthase Transferase 0.6515893
3 (d)CMP-kinase Transferase 0.9362530
4 (d)CMP-kinase Transferase 0.9362530
5 (R)-citramalate-synthase Transferase 0.5533522
6 (R)-citramalate-synthase Transferase 0.5533522The Awesome Procedures for Neo4j (APOC) library extends user options for data analysis and conversion beyond the limitations of CYPHER.
For more information can be found at the APOC github repository.