High-performance method for identification of super enhancers from ChIP-Seq data with configurable cloud virtual machines

A universal method for rapid identifying super-enhancers which are large domains of multiple closely-spaced enhancers is proposed. The method applies configurable cloud virtual machines (cVMs) and the rank-ordering of super-enhancers (ROSE) algorithm. To identify super-enhancers a сVM-based analysis of the ChIP-seq binding patterns of the active enhancer-associated mark is employed. The use of the proposed method is described step-by-step: configuration of cVM; ChIP-seq data alignment; peak calling; ROSE algorithm; interpretation of the results on a client machine. The method was validated for the search of super-enhancers using the H3K27ac mark in the sample datasets of a cell line (human MCF-7), mouse tissue (heart), and human tissue (adrenal gland). The total analysis cycle time of raw ChIP-seq data ranges from 15 to 48 min, depending on the number of initial short reads. Depending on the data processing step and availability of multi-threading, a cVM can be scaled up to a multi-CPU configuration with large amount of RAM. An important feature of the method is that it can run on a client machine that has low-performance with virtually any OS. The proposed method allows for simultaneous and independent processing of different sample datasets on multiple clones of a single cVM. • Cloud VMs were used for rapid processing of ChIP-seq data to identify super-enhancers. • The method can use a low-performance computer with virtually any OS on it. • It can be scaled up for parallel processing of individual sample datasets on their own VMs for rapid high-throughput processing.