Sevastianos Korsak
supervisor: Dariusz Plewczynski
In this poster we discuss how we can create simulations from stripes in ChIA-PET datasets. Stripes usually are connected with one-sided loop extrusion (LE), which is a biological process mediated by two factors: an Smc complex like cohesin or condensin which has a ring-like structure and acts as motor so as as to extrude the loop, and boundary elements like CTCF proteins which act as barriers for Smc complexes and they indicate the ending points of LE mechanism. So our process can be spitted up in two main steps: (1) the detection of stripes in ChIA-PET data, and (2) having gathered the positions of stripes and the intensities that correspond to them, to create realistic 3D models. For stripes detection we used a computer vision technique named StripeNN, which uses canny edge detection. For the 3d modelling we used OpenMM, which is a library for molecular simulation based on C, which can be called from python.