carajillu/adc_developability

Workflows used in the paper "Something Something Developability of Antibody-Drug Conjugates"

Jupyter Notebook

0

65 commits

updated Aug 26, 2026

See the code

README

adc_developability

Workflows used in the paper "Something Something Developability of Antibody-Drug Conjugates"

Installation

Beware, this is a "works on my machine" installation guide, writing it for myself so that I know what I've done.

Cloning the repo

cd $GENERIC_SOFTWARE_FOLDER # this is ~/github for me
git clone https://github.com/carajillu/adc_developability.git
cd adc_developability

Environment

It is recommended to create a .venv or .conda environment to run these workflows:

.venv

python -m venv .venv
source .venv/bin/activate # version will be whichever python version you have installed

.conda

conda create --prefix $(pwd)/.conda python=3.10 # with conda you can specify version which might be handy
conda activate $(pwd)/.conda

Requirements

  • pyTorch (torch) installed via pyproject.toml. Version TBC
  • transformers installed via pyproject.toml. Version TBC

redim

This is the dimensionality reduction package I wrote to make things easier for myself. To install

cd $GENERIC_SOFTWARE_FOLDER # this is ~/github for me
git clone https://gitlab.cis.strath.ac.uk/ixb26106/redim.git
cd redim
pip install . # dependency versions should be pretty standard

Installation

cd $GENERIC_SOFTWARE_FOLDER # this is ~/github for me
cd adc_developability
pip install .

Running the pipeline

I will first write the pipeline, then explain here how to run it

carajillu/adc_developability

Workflows used in the paper "Something Something Developability of Antibody-Drug Conjugates"

Jupyter Notebook

0

65 commits

updated Aug 26, 2026

See the code

README

adc_developability

Workflows used in the paper "Something Something Developability of Antibody-Drug Conjugates"

Installation

Beware, this is a "works on my machine" installation guide, writing it for myself so that I know what I've done.

Cloning the repo

cd $GENERIC_SOFTWARE_FOLDER # this is ~/github for me
git clone https://github.com/carajillu/adc_developability.git
cd adc_developability

Environment

It is recommended to create a .venv or .conda environment to run these workflows:

.venv

python -m venv .venv
source .venv/bin/activate # version will be whichever python version you have installed

.conda

conda create --prefix $(pwd)/.conda python=3.10 # with conda you can specify version which might be handy
conda activate $(pwd)/.conda

Requirements

  • pyTorch (torch) installed via pyproject.toml. Version TBC
  • transformers installed via pyproject.toml. Version TBC

redim

This is the dimensionality reduction package I wrote to make things easier for myself. To install

cd $GENERIC_SOFTWARE_FOLDER # this is ~/github for me
git clone https://gitlab.cis.strath.ac.uk/ixb26106/redim.git
cd redim
pip install . # dependency versions should be pretty standard

Installation

cd $GENERIC_SOFTWARE_FOLDER # this is ~/github for me
cd adc_developability
pip install .

Running the pipeline

I will first write the pipeline, then explain here how to run it