Skip to main content
Fig. 15 | BioData Mining

Fig. 15

From: ParticleChromo3D: a Particle Swarm Optimization algorithm for chromosome 3D structure prediction from Hi-C data

Fig. 15

Loss Function comparison without random XYZ coordinate initialization. We used the same XYZ coordinates for the initialization across the loss function run execution for each of the chromosomes 1,10,12,19, and 21 of 1 MB GM12878 cell Hi-C data. With this configuration, we had the same initialization for the algorithm to use for the 3D structure reconstruction for the loss functions SSE, RMSE, Huber-0.9, Huber -0.5, Huber-0.1, and MSE. The Y-axis denotes the DSCC metric score in the range [-1,1], and X-axis denotes the chromosomes. A higher DSCC value is better

Back to article page