theislab / cellrank2_reproducibility Goto Github PK
View Code? Open in Web Editor NEWCellRank 2's reproducibility repository.
License: BSD 3-Clause "New" or "Revised" License
CellRank 2's reproducibility repository.
License: BSD 3-Clause "New" or "Revised" License
The scEU-seq data needs to be analysed using dynamo. This includes velocity information, fixed point analysis, and least action path analysis.
Using plot_states
to save a PDF produces an unreadable file due to dpi
being specified. Instead of using dpi=400
as a default, it should be set to matplotlib's "figure"
.
To adjust the DPI value used when saving figures, dpi
needs to be added as an argument to plot_states
.
Add GitHub actions CI to check linting of scripts and notebooks.
The README of the repository and index page of the rendered jupyter book are currently out of sync after #43.
We need to add the reproducibility notebooks for the NeurIPS analysis using the PseudotimeKernel.
Add TSI metric for each dataset analysed.
cytotrace_vs_cr2_data_perparation.ipynb
as the notebook contains both data preparation and the actual score analysisTo compare macrostates (terminal states) based on different kernels, the purity w.r.t the clusters based on which they are annotated can be used. To this end, we need one function to compute the state purity, and one function to visualize it.
Add notebooks to identify initial states in datasets instead of terminal states.
Set up generic project structure with pre-commit hooks, pyproject.toml
file, directories for data, figures, notebooks, README, etc.
ATM, the data/
and figures/
directory are note ignored by git. The corresponding lines are commented out in .gitignore
.
Compare the CR2 kernels to an RNA velocity-based workflow.
When installing cellrank/cellrank2 using conda-forge, even following the directions from your repository with:
conda install -c conda-forge cellrank
I get an ImportError .... : undefined symbol: H5Pget_fapl_direct:
error when trying import cellrank as cr
in my Jupyter notebook.
However, if I install everything in pip, there's no issues at all.
Thanks!
Remove the plot_states
function and rely on the estimator method plot_macrostates
, instead, after it has been updated to match plot_states
.
Update colors used to compare methods for analysing the scEU-seq intestinal organoid data.
To easily check how a given set of gene ranks among drivers towards a macrostate, a corresponding function is needed. The function can/should also be able to quantify how many of the provided genes rank below a given threshold (e.g. among the top 100 drivers).
We need to add the reproducibility notebooks to
We need to add the reproducibility notebooks
The column "obs_col"
is defined multiple times.
To compute a real-time-informed pseudotime, the transition matrix needs to be symmetrized to apply DPT on it. A corresponding helper function get_symmetric_transition_matrix
needs to be implemented.
plot_macrostate
function directly instead of custom plot_states
.pathlib
to make directories instead of os
.FIGURE_FORMAT
.To plot state (macrostates, terminal states) purity with a custom order, the plot_state_purity
function needs to be updated/generalized.
To analyze gene trends, a pseudotime needs to be constructed for the scEU-seq data.
To easily and comprehensively visualize identified macrostates or terminal states, they can be plotted with an outline on a given embedding. To do so (and reduce duplicate code), a corresponding plotting function needs to be implemented.
Add or update notebooks to highlight TSI and CBC scores.
We need to add the reproducibility notebooks to
The currently rendered Jupyter book of the repo does not include the notebooks of the hematopoiesis and embryoid body development analysis. The issue are incorrect paths to the respective files.
Imports from cr2
are not possible because of an ill-defined __init__.py
file.
from cr2 import running_in_notebook
File [/cellrank2_reproducibility/src/cr2/__init__.py:1]
----> 1 from utils import running_in_notebook
3 __all__ = ["running_in_notebook"]
ModuleNotFoundError: No module named 'utils'
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.