Description

This repo provides code for lightsheet imaging data analysis pipeline for extraction of quantitative readouts regarding vascular volume and drug penetration in tumors.

A custom data analysis pipeline was developed to enable rapid analysis of tumor lightsheet datasets. Key goals of this analysis pipeline are:

• Enable extraction of quantitative readouts regarding drug penetration in whole tumors from lightsheet data sets. See Dobosz et al. 2014., Neoplasia¹ for reference.

• Use python programming to make use of open source packages that can support building a custom pipeline.

• Use cloud computing environment (Merck High Performance Computing Resources) to enable rapid analysis of very large lightsheet data sets.

• Develop an analysis pipeline with two run modes:

  a) Automatic run mode: enables executing full analysis pipeline on a new data set via a single line of code

  b) Lego Brick mode: enables re-using parts of the analysis pipeline for building new analysis methods.

Code Documentation

The code documentation and examples may be found here

Learn more in our publication:

---

Title: A Light sheet fluorescence microscopy and machine learning-based approach to investigate drug and biomarker distribution in whole organs and tumors.

***

Authors:

Niyanta Kumar, Petr Hrobař, Martin Vagenknecht, Jindrich Soukup, Nadia Patterson, Peter Bloomingdale, Tomoko Freshwater, Sophia Bardehle, Roman Peter, Ruban Mangadu, Cinthia V. Pastuskovas, and Mark T. Cancilla

Merck & Co., Inc.

---

Development Team

In case of need of any editional (technical) information reach out to the development team via github's issue:

• Martin Vagenknech (Merck & Co., Inc.)
• Petr Hrobar (Merck & Co., Inc.)
• Jindrich Soukup (Merck & Co., Inc.)

---

Installation Process

Repo is a python package. Installation process can be automated via bash .sh file in the environment_setup folder.

1) Code Download

Clone the github repository (Entire project in one folder) by running:

# Clone The repo localy to your computer
git clone https://github.com/Merck/3D_Tumor_Lightsheet_Analysis_Pipeline.git

# Navigate to the repo folder
cd 3D_Tumor_Lightsheet_Analysis_Pipeline

2) Dependencies Installation:

When installing the package:

• MAC/LINUX Users

  2.1) Make sure you have conda installed on your computer. if not, you may use this link.

  2.2) Create a python environment
  Run in the terminal:

  source environment_setup/set_3d_infrastructure.sh

• Windows Users

  2.1) Make sure you have conda installed on your computer. if not, you may use this link.

  2.2) Create a python environment run all lines of environment_setup/set_3d_infrastructure.sh manually in the terminal

Hardware Requirements:

To operate the code on local computers we recommend the following MINIMAL Hardware Requirements:

• CPU with at least 6 Cores
• 16 GB RAM
• 800 GB Storage for the Data
• GPU is only required when deep learning model (UNET) is being used.

Copyright

Copyright © 2022 Merck & Co., Inc., Kenilworth, NJ, USA and its affiliates. All rights reserved.

Refs

Footnotes

1. Dobosz, M., Ntziachristos, V., Scheuer, W. & Strobel, S. Multispectral Fluorescence Ultramicroscopy: Three-Dimensional Visualization and Automatic Quantification of Tumor Morphology, Drug Penetration, and Antiangiogenic Treatment Response. Neoplasia 16, 1-U24, doi:10.1593/neo.131848 (2014).* ↩