ADICyt is a desktop software application for effective analysis of flow cytometry data. Great emphasis is placed on leveraging machine learning methods for automated data analysis and population discovery(gating), to speed up and improve analysis of newly gathered measurements using statistical algorithms. The software uses its own algorithm developed by ADINIS Ltd., that has ranked first in all participated categories in the FlowCap international challenge in competition with multiple top US and Canadian scientific labs.
The heart of ADICyt’s automated populations identification lies in mixture model based clustering. Mixtures of the model follow Student’s-t distributions. The clustering process can be run in two modes. The first one is multidimensional, where all the dimensions are considered at once in a single run of the clustering algorithm to discover cell populations. The second mode is iterative. The algorithm hierarchically (top to bottom) splits the data to a sequence of manual gating events, mimicking a human expert. At any level of the hierarchical splitting, the algorithm tries to identify the “optimal” 2D-projection that differentiates the data to subpopulations.
The performance critical parts are implemented to run on NVIDIA GPU using CUDA C technology, which dramatically improves speed of the algorithm. If appropriate graphic card is not present in the system, all code will be run on CPU utilizing multiple cores if present. In the case of 80000 cells, a single clustering run on two dimensions takes on GPU ( NVIDIA GeForce GTX 285 ) about 13 seconds while on 4 CPU cores (Intel® Core™ i5-750 ) about 46 seconds.
ADICyt can run both locally and in server mode where computationally intensive parts of algorithm run on server.
ADICyt was developed in cooperation with the Cancer Research Institute of the Slovak Academy of Sciences.
ADICyt – demonstration video
Obtain your free FULL version
of ADICyt. Download now:
To obtain the free full license key, please send us your hardware key to firstname.lastname@example.org.
- ADICyt for automated clustering of flow cytometry data
- ADprot for automated search in protein databases
- ExProf for assembly and comparison of expression profiles
- InDelFinder for automated search of mutations in DNA sequences
- Logit for automatic detection of data dependencies by means of logistic regression
- MultipluginG for extending NextGen sequencing funcionality of Geneious software
- Parent Prophet for testing of custom relationships between people