Sagittarius-api 0.17

No project description provided

Sagittarius

Gene expression time-series extrapolation for heterogeneous data

Introduction

Sagittarius is a model for temporal gene expression extrapolation simulate unmeasured gene expression data from unaligned, heterogeneous time series data. This is a python repository to simulate transcriptomic profiles at time points outside of the range of time points available in the measured data.

https://github.com/addiewc/Sagittarius

Installation Tutorial

pip install Sagittarius-api

System Requirements

Sagittarius is implemented using Python 3.9 on LINUX. Sagittairus expects torch==1.9.1+cu11.1, numpy==1.21.2, pandas==1.3.3, scikit-learn=0.24.2, matplotlib==3.4.3, seaborn==0.11.2, umap-learn=0.5.1, anndata=0.8.0, statsmodels==0.13.0, tqdm==4.62.3, and so on. For best performance, Sagittarius can be run on a GPU. However, all experiments can also be run on a CPU by not setting the --gpu flag. Typical installation requires approximately 5 minutes.

How to use our code

# Use pretrained_model
>>> from Sagittarius import simulate_measurements_webserver as smw
>>> smw.EvoDevoSimulation('Chicken', 'Heart', 2.0).head()
"""
            DPM1     GCLC      NFYA    NIPAL3     WNT16      ICA1    DBNDD1     ALS2     CFLAR     TFPI  ...      MRC1       GAN     MMP12   OTUD7B    STRADA     NCOA4   RASL10B    MMP28    H0YAA0    GRIN2B
timepoint                                                                                                   ...                                                                                                  
2.0        1.583706  1.51412  1.448546  1.438742  0.432965  1.305412  1.182693  1.27424  1.346429  1.37519  ...  0.369062  0.726691 -0.025032  1.20409  1.493279  1.722517  1.391992  1.04687  1.558734  0.317554
"""

# Use model and config
>>> from Sagittarius import simulate_measurements
>>> model_path = 'model.pth'
>>> config_file = 'cfg.json'
>>> species = ['chicken', 'human']
>>> organ = ['brain', 'cerebellum']
>>> timepoint = 2.0
>>> adata_res = simulate_measurements.simulate_single_EvoDevo(
    model_path, config_file, species, organ, timepoint)