Cellphone DB (and other database versions)

[SimonDMurray]

Hi,

Sorry if this information is available somewhere but I cannot find it. I would like to know what version of cellphoneDB (and other databsases) you have inside the liana-py package.

I am aware you import your databases from OmniPathR (unless I am mistaken) and have checked their version of cellphoneDB: saezlab/OmnipathR#85

I wanted to double check that your version matches theirs and if not could it be updated?

Thanks,
Simon

[dbdimitrov]

Hi @SimonDMurray,

Currently, the CellPhoneDB accessible via LIANA will be CellPhoneDBv2, as while it's imported via OmniPath, the database in LIANA+ is versioned independently. If you wish to use CellPhoneDBv4 - you can directly obtain it via OmniPath's python (https://github.com/saezlab/omnipath) or R API and then just feed it to LIANA as a dataframe.

I plan to update it also in LIANA, but this also includes some infrastructural changes which we're currently working on (#9).

[dbdimitrov]

Hi @SimonDMurray,

A bit delayed but I was checking the new CPDB resource and one can get it the following way:

import pandas as pd
import numpy as np

import requests
import io

# read csv from link
# https://github.com/ventolab/cellphonedb-data/blob/master/data/interaction_input.csv
resource = requests.get('https://raw.githubusercontent.com/ventolab/cellphonedb-data/master/data/interaction_input.csv').content
resource = io.StringIO(resource.decode('utf-8'))
resource = pd.read_csv(resource, sep=',')
# keep only PPIs
resource = resource[resource['is_ppi']][['interactors']]
# replace + with _
resource['interactors'] = resource['interactors'].apply(lambda x: x.replace('+', '_'))
# if interactors contains two '-' replace the first one with '&
resource['interactors'] = resource['interactors'].apply(lambda x: x.replace('-', '&', 1) if x.count('-') == 2 else x)
# split by - and expand
resource = resource['interactors'].str.split('-', expand=True)
# replace & with - in the first column
resource[0] = resource[0].apply(lambda x: x.replace('&', '-'))
resource.columns = ['ligand', 'receptor']

Then it's as simple as passing the resource to the resource parameter to any liana function you would like to use.

[feriobrahmana]

Hi @dbdimitrov,

I’m new here and just came across your LIANA+ paper—it seems incredibly useful for my work!

However, I still have a few questions, even after reading this thread:

1. Based on your tutorial and your answer in this discussion, it seems possible to feed LIANA+ with a specific database (e.g., CellChatv2) that isn’t included in OmniPath. Is that correct?
2. If I understand correctly, the key requirement is ensuring the format of the dataframe being fed into LIANA+. Could you clarify how the dataframe should be formatted? I tried searching for this but couldn’t quite figure out what to look for.
3. Lastly, if I use a custom database, can I still utilize the rank-aggregation method to analyze results across several methods in parallel?

Thank you so much!

Best,
Ferio

[dbdimitrov]

Hi @feriobrahmana,

Yes, you can feed any database to liana, to see the format you can just do liana.resource.select_resource(). This will return a dataframe, with the columns that it expects. Then you can feed a DataFrame with the same set of columns ('ligand', 'receptor') to the resource parameter of any method. Likewise, you can provide a list of tuples to the interactions parameter; e.g. [('ligand1', 'receptor1_receptor3'), ('ligand2', 'receptor2'), ...].

Note that '_' designates the separation of complex subunits.

Any method in liana will work with these parameters, including rank-aggregation.

Hope this helps.

Daniel

[feriobrahmana]

Ah, okay! I see...

Thank you so much for the swift and clear response, @dbdimitrov!

And, yes, I think your answer has helped a lot!

Will start to try and find a great utilization of LIANA+ then, hehe!

Best,
Ferio