As a result of recent high throughput sequencing efforts a highly complex and diverse population of sRNAs was revealed in the cell. However, as sequencing becomes feasible for multiple samples, it is important to develop methods for analysis of such data and annotation of sRNA loci, i.e. regions of the genome that produce sRNAs in a coordinated manner. The predictions provided by current methods indicate that location on the genome can no longer be considered a discriminative feature between close regions. Thus we present a new method, called CoLIde (Coexpression Loci Identification), to predict sRNA loci from multiple high throughput sequencing datasets. The method is based on both location on the genome and the variation of sRNA expression levels in the samples.