PIQ is a new computational method that uses machine learning to identify the genomic binding sites of hundreds of transcription factors (TFs) at corresponding motifs from DNase-Seq experiments with accuracy comparable to ChIP-Seq. We employed PIQ in a developmental lineage paradigm of pancreatic endoderm specification from embryonic stem cells and identified a set of novel pioneer TFs that dynamically open chromatin in the lineage and enable other TFs to bind to adjacent DNA. Surprisingly, several pioneer TFs with non-palindromic motifs only open chromatin in one direction. We found that genomic binding of a distinct subset of settler TFs depends on proximity to chromatin opened by pioneer TFs. In vivo experiments confirmed directional and non-directional pioneer activity and settler binding.