Change Log

The record of all notable changes to this project will be documented in this file.

v1.3.3

June 16, 2026

  • Add predict_with_pretrained_model(...) as a one-call entry point for Example 1.

  • Add automatic downloading of essential Example 1 assets by default:

    • The pre-trained model_DeSide.h5 is downloaded from Figshare (DOI: 10.6084/m9.figshare.25117862.v1).

    • The remaining DeSide_model/* metadata files and datasets/gene_set/*.gmt files are downloaded from the DeSide_mini_example repository.

  • Add explicit download logs for each downloaded file, including destination path, file size, and md5.

  • Add a runnable script examples/example1_pretrained_model.py for non-notebook usage.

  • Update documentation and README to reflect the new one-line workflow and auto-download behavior.

v1.3.2 (revision)

September 15, 2024

  • Update documentation.

v1.3.1 (revision)

July 13, 2024

  • Update Figure 1ab.

  • Update workflow and plot functions during the revision.

v1.2.2 (revision)

February 3, 2024

  • Update the documentation.

v1.2.0

January 16, 2024

  • Add GEP-level filtering in the PCA space of TCGA samples.

  • Add Dirichlet distribution to simulate the cellular proportions of bulk RNA-seq data.

v1.1.0

October 1, 2023

  • Update the workflow to include:

    • alpha_total_rna_coefficient: coefficient of total RNA for each cell type during bulk RNA-seq simulation.

    • cell_type2subtypes: mapping of cell types to subtypes.

  • Update the DNN model to include pathway profiles.

  • Update the documentation.

v1.0.2 (bioRxiv)

June 2, 2023

  • Update documentation.

v0.9.8

November 19, 2021

  • Build the whole workflow in one file.

  • Record main parameters and running logs.

v0.9.7

September 13, 2021

  • Filter single cell data of CD4 T and CD8 T cells by the ratio of marker genes.

  • Update marker gene list and re-annotate sub-clustering based on dot-plot figures of each single cell dataset.

v0.9.6.1

July 16, 2021

  • Set total_cell_number to 100 and n_base to 3 when simulating bulk cell data to better keep cell type diversity.

v0.9.6

July 7, 2021

  • Fix typos.

  • Check whether Cancer Cells exists and update bbknn to v1.5.1.

v0.9.5

July 2, 2021

  • Fix typos.

  • Update Scanpy to v1.8.0.

  • Update seaborn to v0.11.1.

v0.9.4

June 11, 2021

  • Fix bugs.

  • Update Scanpy to v1.7.2.

  • Evaluate correlation between predicted cell fractions and mean marker-gene expression per cell type.

v0.9.3

May 18, 2021

  • Fix bugs.

  • Add support for more tumor types.

v0.9.2

May 12, 2021

  • Build workflow and create documentation.

v0.2

October 1, 2020

  • First release.