Several years ago, the CRISPR system has emerged into our lives, bringing the promise for curing diseases, improving crops quality, and facilitating laboratory research by editing targeted genes. The modification of specific genes and genomic regions is enabled by designing CRISPR to target them according to sequence similarity to the DNA sequence. However, the efficacy of one design is not uniquely defined by exact identity to the target site, thus additional genes might be unintentionally modified. I present CRISTA (CRISPR Target Assessment), a novel algorithm within the machine learning framework that determines the probability of a genomic site to be cleaved by a given CRISPR design.