Run microbetag from Python

Use microbetag’s features as a library

Assuming you have installed microbetag locally, you can now use microbetag’s features independently.

To go through them, you may have a look at the API Reference but also on the unittests of ours. You may run individual genome annotation steps, build your own GENREs, as well as the routines allowing you to export pathway and/or seed complementarities.

Feel free to contact on our Matrix community in case you face any trouble.

Use microbetag API programmatically

One may use the API routes described through Python. For example, to get the microbetagDB genomes related species with NCBI Taxonomy id 853:

import requests
url = "https://msysbio.gbiomed.kuleuven.be/ncbiTaxId-to-genomeId/853"
r = requests.get(url)
r.status_code
200
r.json()
{"537007": ["GCA_002222595.2"]}

Furthermore, one may run microbetag directly and not through the CytoscapeApp by simply converting their abundance table or co-occurrence network in a JSON object. They would also need to provide the following arguments as part of the JSON obect:

arguments_list = [
  "input_category", 
  "taxonomy", 
  "phenDB", 
  "faprotax", 
  "pathway_complement", 
  "seed_scores"
]

Here is an example using an ASV abundance table as input, asking for all microbetag annotations supported:

import requests
import json

# Init a json object
json_object = {}

# Load the abundance table and after removing any blanks load it as lines in the json object
data = open("my_abundance_table.tsv","r")
data = data.readlines()
data = [line.rstrip("\n").split("\t") for line in data]
json_object["data"] = data

# Likewise, in case you already have a network of your own 
edgelist =  open("edgelist.csv","r")
edgelist = edgelist.readlines()
edgelist = [line.rstrip("\n").split("\t") for line in edgelist]
json_object["network"] = edgelist


# Make a dictionary with your arguments settings
args = {
  "taxonomy":"Silva", 
  "delimiter": ";", 
  "faprotax": True, 
  "phenDB": True, 
  "pathway_complement": True,
  "seed_scores": True,
  "manta": False,
} 
json_object["inputParameters"] = args

# Load your metadata file like in the data case; if no metadata you can skip this
# json_object["metadata"] =

# Set the url to the microbetag server
url = "https://msysbio.gbiomed.kuleuven.be/upload-abundance-table-dev"

# Run microbetag
r = requests.post(url, json = json_object)

# Save your annotated network to a json file that cytoscape can load
response_dict = r.json()
with open('new_microbetag.cx', 'w') as f:
	json.dump(response_dict, f, indent=4, sort_keys=True)

Apparently, the my_abundance_table.tsv and the edgelist.csv files can be in any format initially. Yet, they need to be converted in a way so eventually what you send to the microbetag server is in the form of data and network in the above chunk. These look like this:

>>> data
[
  ['seqId', 'sample1', 'sample2', 'taxonomy'],
  ['bin_100', '77', '1', 'd__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Chitinophagales;f__Chitinophagaceae;g__Terrimonas;s__Terrimonas ferruginea'], ...
]
>>>edgelist
[
  ['nodeA', 'nodeB', 'weight'], 
  ['bin_45', 'bin_28', '0.471'], ...
]

Important

POSSIBLE ARGUMENT’S VALUES

input_category: [abundance_table | network]

taxonomy: [GTDB | Silva | microbetag_prep | other]

phenDB, faprotax, pathway_complement, seed_scores: [True | False], get_children

Warning

You arguments dictionary needs to include all the arguments mentioned above. If any is not provided , microbetag will eventually fail and won’t return an annotated network.

Syntax common error: make sure you do not have a / in the end of the url. If so, microbetag will never start.