About Cytoscape
Cytoscape is an open source tool for visualizing a network with any type of data. With the collaboration of National Resource for Network Biology I developed two dimensional reduction algorithms which work seamlessly with the Cytoscape tool based on clusterMaker2 as a plugin.
My Contribution
Major portion of my project involved developing a Principal Coordinate Analysis and t-Distributed Stochastic Neighbor Embedding in clusterMaker2. Many resources had to be perused in order to identify how those algorithms react in networks. Additionally, estimating the usage of the current API in clusterMaker2 proved to be a challenge.
Based on my mentor’s advice, I completed the algorithm parts before the mid evaluation and connected those algorithms with cytoscape before the final evaluation.
Mentor : Scooter Morris
Work that got merged:
Link contains all the commits to the clusterMaker2 repository.
Snapshots from t-Distributed Stochastic Neighbor Embedding
Issues To Be Solved
The Principal Coordinates Analysis depends on Gower’s matrix. Since it uses a significant amount of memory when processing a large matrix, it should be solved when setting up large matrices as input.
My Experience
GSOC 2k16 was my first ever open source contribution project. Seeing my proposal being accepted was a pleasant surprise to me. Exposure to cutting edge technology and understanding about network based algorithms were significant experiences to my life. My sincere gratitude goes out to my mentor, Scooter Morris, for his invaluable guidance throughout the entire project. His mammoth contribution through thick and through thin was my strength in this achievement.
Mentor's Feedback
