With the increasing number of studies focusing on PIWI-interacting RNA (piRNAs), it is now pertinent to develop efficient tools dedicated towards piRNA analysis. Researchers at the National Institute of Biomedical Genomics, India have developed a novel cluster prediction tool called PILFER (PIrna cLuster FindER), which can accurately predict piRNA clusters from small RNA sequencing data. PILFER is an open source, easy to use tool, and can be executed even on a personal computer with minimum resources. It uses a sliding-window mechanism by integrating the expression of the reads along with the spatial information to predict the piRNA clusters. The researchers have additionally defined a piRNA analysis pipeline incorporating PILFER to detect and annotate piRNAs and their clusters from raw small RNA sequencing data and implemented it on publicly available data from healthy germline and somatic tissues. They compared PILFER with other existing piRNA cluster prediction tools and found it to be statistically more accurate and superior in many aspects such as the robustness of PILFER clusters is higher and memory efficiency is more. Overall, PILFER provides a fast and accurate solution to piRNA cluster prediction.
Flowchart and a block diagram describing cluster prediction using the PILFER algorithm
PILFER is a modular method which can be incorporated into an existing pipeline without much hassle. The major steps in predicting a cluster is shown in the figure with the additional block diagram explaining the peak selection method. Wherever a read satisfies the peak selection criterion, the adjacent reads within 100 kb window are selected to maximize the read count within that cluster. The red boxes represents reads with their corresponding count value inside it which were included in a cluster whereas the purple boxes represent reads which do not form a part of any cluster.