Genome-wide Prediction of Potential Polycomb Response Elements and their Functions

Abstract

The Polycomb-group proteins (PcG) and Trithorax-group proteins (TrxG) are two major epigenetic regulators important for proper differentiation during development (1,2). In Drosophila melanogaster (D. melanogaster), Polycomb response elements (PREs) are short segments of DNA with a high density of binding sites for transcription factors (TFs) that recruit PcG and TrxG proteins to chromatin. Each PRE has a different number of binding sites for PcG and TrxG, and these binding sites have different topological organizations. It is thus difficult to find general rules to discover the locations of PREs over the entire genome. We have developed a framework to predict the locations and roles of potential PRE regions over the entire D. melanogaster genome using machine learning algorithms. Using a combination of motif-based and simple sequence-based features, we were able to train a random forest (RF) model with very high performance in predicting active PRE regions. This model could distinguish potential PRE regions from non-PRE regions (precision and recall ~0.92 upon cross-validation). In the process, the model suggests that previously unrecognized TFs might contribute to PcG/TrxG recruitment at the PRE locations, as the presence of binding sites for those factors is strongly informative of active PREs. A secondary regression model provides information on features that further differentiate PREs into functional subclasses. Our findings provide both new predictions of 7887 potential PREs in the D. melanogaster genome, and new mechanistic insight into the set of DNA-associated proteins that may contribute to PcG recruitment and/or activity. Author summaryDuring the development of multicellular organisms, the pattern of gene expression for every cell type must be established and then inherited through cell division. DNA cis-regulatory elements called Polycomb response elements (PREs) are major drivers of the spatio-temporal regulation of many genes during development. Recently, it has been reported that, depending on the cell type and condition, these elements could have dual functionality: a PRE can act as a silencing element in one cell type and an enhancer in others. Integrating all binding information of characterized TFs with DNA sequence features allowed us to construct a highly informative machine learning model to predict and classify PRE locations. Applying our model to Drosophila both allowed us to identify many new putative PREs (15-16 times the number currently known from experimental studies) and provides insight into the specific DNA-binding proteins that may additionally contribute to PRE function.

Persistent epigenetic reprogramming of sweet taste by diet

“…Identification of predicted PRE sites was performed exactly as described in (28), using the dm6 genome. As suggested in (28), we use a threshold confidence score of 0.8 to identify the PREs used in the present analysis. A complete list of predicted PREs, with accompanying confidence scores, is shown in file S1.…”

“…To identify likely targets of each transcription factor of interest, we drew upon the transcription factor binding site calculations described in (28) in which the motif of each transcription factor was scanned along every base pair of the D. melanogaster dm6 genome using FIMO (71) and the base-pairwise binding results were converted to robust z scores. For each TF, we then considered its regulon to consist of all genes with at least one binding site with z score above a TF-specific threshold within 2-kb upstream of the beginning of the gene.…”

Persistent Epigenetic Reprogramming of Sweet Taste by Diet

“…Response Elements (PREs), cis-regulatory sequences to which Polycomb Group Proteins bind in D. melanogaster (Chan et al, 1994) (Müller and Kassis, 2006) . Using a recently developed tool that predicts PRE regions genome-wide (Khabiri and Freddolino, 2019) , we found that Pcl was present in these regions ( Fig. 3D To determine changes in Pcl occupancy with diet, we compared Pcl peaks between flies fed a control and sugar diet.…”

“…Identification of predicted PRE sites was performed exactly as described in (Khabiri and Freddolino, 2019) , using the dm6 genome. As suggested in (Khabiri and Freddolino, 2019) , we use a threshold confidence score of 0.8 to identify the PREs used in the present analysis. A complete list of predicted PREs, with accompanying confidence scores, is shown in supplementary file 1.…”

“…To identify likely targets of each transcription factor of interest, we drew upon the transcription factor binding site calculations described in (Khabiri and Freddolino, 2019) , in which the motif of each transcription factor was scanned along every base pair of the D. melanogaster dm6 genome using FIMO (Grant et al, 2011) , and the base pair-wise binding results converted to robust z-scores. For each TF, we then considered its regulon to consist of all genes with at least one binding site with z-score above a TF-specific threshold within 2 kb upstream of the beginning of the gene.…”