mibilib

Python client for IONpath MIBItracker API, plus utility functions for working with MIBItiff images.

https://ionpath.github.io/mibilib/

mibilib has been updated for compatibility with MIBIscope/mibin-commercial MIBItiff images.
In particular, custom HIMSR tags have been added to the existing ImageDescription TIFF tag:

• himsr.author: @christianrickert
• himsr.back: level of background correction (gold slide interference)
• himsr.bloom: level of bloom correction (horizontal stripes)
• himsr.creation: timestamp at time of MIBItiff creation
• himsr.mph: median pulse height for normalization
• himsr.software: "mibin-commercial, v.0.9.0"
• himsr.window: mass window for target

Optional tags can be added to the metadata:

• mibi.norm: normalization factor

Prefix optional tags with mibi. for compatibility with Bio-Formats.

Setup

Install Miniconda

Install the lastest version of Miniconda. Even if your system already has a standalone version of Python installed, it is strongly recommended to use the conda environment manager to install and manage this library's dependencies.

Install mibilib with conda and pip

conda env create -f environment.yml
conda activate mibilib

Usage

It's easy to work with MIBITIFFs using the mibidata module.

from mibidata import tiff

# extract metadata from mibitiff (fast)
metadata = tiff.info("mibi.tif")  # returns a dictionary

# print FoV name from metadata
print(metadata["fov_name"])

# read a mibitiff into a MibiImage class instance (slow)
mibitiff = tiff.read("mibi.tif")

# print channel tuples from mibitiff
print(mibitiff.channels)

# copy MibiImage class into a new instance
newtiff = mibitiff.copy()

# rename a target channel
OLD_TARGET = "dsDNA"
NEW_TARGET = "dsDNA (nuclear marker)"
newtiff.rename_targets({OLD_TARGET: NEW_TARGET})

# normalize pixel data of all channels
NRM = 1.23
newtiff.data *= NRM

# offset data of second channel only
CHANNEL = 1
newtiff.data[:, :, CHANNEL] += 3.0

# add custom tags with a dictionary
TAGS = {"mibi.norm": NRM}
newtiff.add_attr(**TAGS)

# write MibiImage class into a file
tiff.write("new.tiff", newtiff)

To use the MIBItracker API with mibitracker, you will need to use the backend url listed in the About page. This can be accessed after you have logged in from the menu under your username in the upper right of the window.

from mibitracker.request_helpers import MibiRequests

request = MibiRequests(
    'https://[sitename].api.ionpath.com',
    '[email protected]',
    'password1234'
)
image_id = request.image_id('20180927', 'Point3')
image_details = request.get('/images/{}/'.format(image_id))

More examples can be found in the following notebooks:

• MIBItracker_API_Tutorial

• MibiImage_Tutorial

• SingleCellSpatialExamples

Full documentation for this library can be found at https://ionpath.github.io/mibilib/.

Sample data

Access to sample data to run the tutorials in the notebooks can be requested by creating an account at the following URL: https://mibi-share.ionpath.com.