Motif prediction in nucleotide sequences using Markov and hybrid models

Transcription factor binding to DNA is a fundamental mechanism of gene regulation. These proteins recognize specific nucleotide sequences, but several other effects coregulate the process. Here, we describe and compare various motif-finding and motif-description methods. We then apply energy-based approaches to characterize transcription factor binding motifs of methylation-sensitive (or non-sensitive) transcription factors. Next, we employ the deep-learning model Enformer to investigate how different transcription factors contribute to promoter-enhancer communication. Finally, we examine the influence of motif context, proposing a role in directing transcription factors to their target sites. 1