top of page
Search
vitod1

AN APPLICATION OF LATENT CLASS MODELS FOR DICHOTOMOUS OBSERVATIONS TO EXPERIMENTAL DATA FROM TRANSCR

Alessandro Testori, PhD, IRGB-CNR Anna Maristella Steri, PhD, IRGB-CNR


Poster Not on Display


Transcription Factors (TFs) bind DNA at different locations in the human genome and typically do not act independently, but tend to assemble, forming patterns. Therefore, it is plausible that multiple binding sites group into several classes, each characterized by specific combinations of TFs. Transposable Elements (TEs) are mobile elements that potentially have the ability to move across the genome and spread regulatory sequences. Most TEs are immobile today; however, over the course of evolution, they have acted as potential sources of binding motifs for TFs. An open question is whether TEs played a role in creating binding sites for TF patterns. Using public ChIP-seq data and focusing on a subgroup of the nuclear receptor superfamily - the Steroid Receptors (SRs) - we first retrieved their binding positions in MCF-7 cells, and then scanned these genomic locations for the presence of other TFs and TEs. We applied a latent class model to identify classes of genomic loci that harbor similar patterns of TFs, considering the simultaneous binding information of available TFs at the binding sites for SRs as response variables, and TEs as covariates to assess whether their presence was associated with the identified latent classes. The first latent class was composed of binding sites typically harboring no TF other than SRs. The presence of a TE at a binding location resulted in general in a tendency toward the first latent class, but in few cases (such as LTR/ERV1-type TEs), a significant tendency toward a different latent class was observed. Overall, TEs did not largely contribute to the creation of SR binding site classes characterized by specific combinations of TFs, except in few cases: LTR/ERV1 TEs, which are assumed to have played a central role in the evolution of the SR binding repertoire and in shaping its distinctive TF pattern heterogeneity.

24 views0 comments

Recent Posts

See All

Comments


bottom of page