Utilizing the transfer entropy (TE), TENET successfully reconstructs gene regulatory networks from a range of single-cell RNA sequencing (scRNAseq) data. TENET’s lengthy processing time for big datasets is a drawback, though. We suggest FastTENET, an array-computing variant of the TENET method tuned for acceleration on manycore processors like GPUs, to overcome this constraint. FastTENET uses array computing to calculate the TE by counting the distinct patterns of joint events. FastTENET delivers a performance boost of up to 973× over TENET.
You may find FastTENET at https://github.com/cxinsys/fasttenet.