Fast Priority: A strategy to reduce processing time in a motif discovery tool like Priority

Abstract

Motif recognition is key to identify common sequences and specific elements recognized by different proteins such as Transcription Factors or RNA Binding Proteins. One interesting approach to computationally identify motifs consists on using additional sources of information besides the input sequences to guide the motif search. However, depending on the information and input data size the motif recovery process can take several hours. One of the most used algorithms for motif recovery is Priority [1]. This algorithm uses a Gibbs sampling strategy and includes additional information, in informative priors, to guide the motif search. We propose to reduce motif recovery processing time by using threads to execute independent processes in parallel. We apply the strategy to Priority and the results show the strategy has a positive effect on time and it does not affect the quality of the resultant motif. Reducing processing time in motif recovery algorithms is important considering that massive parallel sequencing is frequently used today. By using massive sequencing the number of sequences derived by experimental techniques is only getting bigger.