LR-DNase: Predicting TF binding from DNase-seq data

ABSTRACTTranscription factors play a key role in the regulation of gene expression. Hypersensitivity to DNase I cleavage has long been used to gauge the accessibility of genomic DNA for transcription factor binding and as an indicator of regulatory genomic locations. An increasing amount of ChIP-seq data on a large number of TFs is being generated, mostly in a small number of cell types. DNase-seq data are being produced for hundreds of cell types. We aimed to develop a computational method that could combine ChIP-seq and DNase-seq data to predict TF binding sites in a wide variety of cell types. We trained and tested a logistic regression model, called LR-DNase, to predict binding sites for a specific TF using seven features derived from DNase-seq and genomic sequence. We calculated the area under the precision-recall curve at a false discovery rate cutoff of 0.5 for the LR-DNase model, a number of logistic regression models with fewer features, and several existing state-of-the-art TF binding prediction methods. The LR-DNase model outperformed existing unsupervised and supervised methods. Additionally, for many TFs, a model that uses only two features, DNase-seq reads and motif score, was sufficient to match the performance of the best existing methods.

Exonic Splicing Enhancer Motif Recognized by Human SC35 under Splicing Conditions

ABSTRACT Exonic splicing enhancers (ESEs) are important ciselements required for exon inclusion. Using an in vitro functional selection and amplification procedure, we have identified a novel ESE motif recognized by the human SR protein SC35 under splicing conditions. The selected sequences are functional and specific: they promote splicing in nuclear extract or in S100 extract complemented by SC35 but not by SF2/ASF. They can also function in a different exonic context from the one used for the selection procedure. The selected sequences share one or two close matches to a short and highly degenerate octamer consensus, GRYYcSYR. A score matrix was generated from the selected sequences according to the nucleotide frequency at each position of their best match to the consensus motif. The SC35 score matrix, along with our previously reported SF2/ASF score matrix, was used to search the sequences of two well-characterized splicing substrates derived from the mouse immunoglobulin M (IgM) and human immunodeficiency virus tat genes. Multiple SC35 high-score motifs, but only two widely separated SF2/ASF motifs, were found in the IgM C4 exon, which can be spliced in S100 extract complemented by SC35. In contrast, multiple high-score motifs for both SF2/ASF and SC35 were found in a variant of the Tat T3 exon (lacking an SC35-specific silencer) whose splicing can be complemented by either SF2/ASF or SC35. The motif score matrix can help locate SC35-specific enhancers in natural exon sequences.

Cell transformation by kFGF requires secretion but not glycosylation.

The Kfgf gene, which encodes a member of the fibroblast growth factor family, was originally discovered by assaying human tumor DNA for dominantly transforming oncogenes. The 22-kD kFGF product contains a single site for asparagine-linked glycosylation and an amino-terminal signal peptide for vectorial synthesis into the endoplasmic reticulum and eventual secretion. To determine whether these features are necessary for transformation, we have constructed mutants of kFGF that are impaired for glycosylation or secretion. All mutants retained the ability to induce DNA synthesis when added to quiescent cells, and the absence of glycosylation had no appreciable effect on the transformation efficiency on NIH3T3 cells. In contrast, mutants of kFGF that remain in the cytoplasm or are retained in the secretory pathway, through addition of a KDEL motif, score negative in standard transformation assays. Since transformation by either the glycosylated or unglycosylated form of kFGF can be reversed by addition of suramin, the data imply that secretion of kFGF, or surface localization of the ligand/receptor complex, is a prerequisite for transformation.