This is the repository for the DisCo source code. The repository contains the source code, install instructions for the necessary dependencies, and compilation instructions.
The code in the presented form was developed specifically for the purpose of deterministically co-encapsulating two particles in droplets on a microfluidic system (as described in Bues, Biocanin, Pezoldt et al. 2020, BioRxiv (https://doi.org/10.1101/2020.05.19.103812). The code base was written in C++ (for performance and latency reasons) and is non-object-oriented.
Specifically, it controls the coordination of real-time analysis of images streamed from a camera connected to a microscope, with the actuation of solenoid valves and pressure regulators connected to a microfluidic device. Detection in a defined ROI is achieved by the following processing sequence 1. subsequent image subtraction, 2. thresholding, 3. hole filling, 4. contour detection, and 5. contour thresholding. Each of the two channels is continuously monitored in this manner to detect and place particles. Particles are detected and stopped by the following sequence of events: 1. Detect particles in a first ROI, 2. Initiate a valve placement sequence (valve oscillation), 3. detect the arrival of the particle in the second ROI. Each channel is controlled by an independent thread, which in contains a thread-pool of detection threads. Both channels are coordinating each other, and are stopping particles subsequently. While concurrent particle stopping is possible, we found that it leads to non predictable states. Upon successful particle placement in both channels a globally coordinated routine is ejecting the particle into a droplet, which is finally captured. An important aspect of the code is elaborate timing of events (apparent by the excessive use of sleep statements), only with which robust working of the system is possible.
This repository contains a Visual Studio solution, which in turn contains the following files:
-
DisCo_main.cpp - Entry point and global control of the UI elements, and channel control.
-
DisCo_functions.cpp - A collection of functions for e.g. channel control, particle detection, encapsulation, etc.
-
DisCo.h - Most important Default definitions of regularly changed variables, control over compiler instructions for calibration menus. Furthermore, includes, function prototypes, global variables.
-
ValveController.cpp - Wrapper class for the NIDAQmx library.
-
ValveController.h - Corresponding header file for corresponding .cpp file.
Furthermore we added a calibration file calib for the OB1 pressure regulator. Please consider this an example file and generate a custom file as described by the manufacturer.
Multiple propriatry libraries, for which we do not hold any distribution rights, are utilized in this project, yet not contained in this repository. These libraries have to be obtained and installed from third party suppliers. Instructions on how to obtain and link said libraries are given below.
This project was generated with Visual Studio (version 2019) on Windows 10. These instructions assume the use of the same by the user. If not installed by default it is necessary to install VS2019 with the C++ toolchain.
Multiple external libraries are utilized in the DisCo source code. As some of these libraries are proprietary they are not distributed with the DisCo source code and need to be installed by the user. In order to compile the DisCo source code from source the following dependencies are required:
- OpenCV 4 (prebuilt libraries are sufficient)
- Ximea xiAPI
- National Instruments NIDAQmx
- Elveflow SDK
The OpenCV 4 prebuilt libraries can be downloaded from the following website: https://opencv.org/releases/
The Elveflow SDK can be downloaded from the following website: https://www.elveflow.com/microfluidic-products/microfluidics-software/elveflow-software-sdk/
Extract the archive, and extract the zipped SDK contained within.
Installer for the NIDAQmx library is located here: https://www.ni.com/en-us/support/downloads/drivers/download.ni-daqmx.html#409845
The installer for the Ximea xiAPI can be downloaded from the following website: https://www.ximea.com/support/wiki/apis/XIMEA_Windows_Software_Package
Note: “…” marks installation specific paths and have to be replaced with the actual installation paths by the user.
- Right click on project select Properties.
- In the Properties menu select correct Configuration: Release and Platform: x64
- Select C/C++ > General. In the field Additional Include Directories select:
- The OpenCV include folder: “…\build\include”
- Elveflow SDK include (example path, version specific): “…\SDK_V3_05_03\DLL64\DLL64”
- Ximea API include: “…\API\xiAPI”
- National Instruments NIDaqMX: “…\Shared\ExternalCompilerSupport\C\include”
- Press approve.
- Go to Linker > General. In field Additional Library Directories add:
- Elveflow SDK library (example path, version specific): …\SDK_V3_05_03\DLL64\DLL64
- Ximea API library: …\API\64bit
- The OpenCV library folder: …\build\x64\vc15\lib
- National Instruments NIDaqMX: …\Shared\ExternalCompilerSupport\C\lib64\msvc
- Go to Linker > Input. In field Additional Dependencies add:
- opencv_world410.lib (Version dependent)
- xiapi64.lib
- Elveflow64.lib
- NIDAQmx.lib
The Visual Studio solution should now contain all necessary dependencies and be able to compile the DisCo executable. Make sure to select Release and x64 for the build.
The main intention of this repository is to provide a compilation ready version of the DisCo code to enable other laboratories to replicate our system. As indicated above, the code represents a specialized solution to the application it was written for, and the focus during development was not on reusability. Hence, although parts of the code could be readily incorporated in derivate projects, we recommend to consider refactoring the code before utilizing it as a development basis.