| Transcription regulation is a fundamental process in biology. Transcription factors are specific proteins which play a key role in transcription regulation by binding to specific regions of the genome. Transcription regulation is highly complex, with multiple transcription factors controlling transcription in a combinatorial manner. Understanding this combinatorial mechanism is critical in understanding the functioning of the cell.;Unfortunately, genomic data describing transcription factor activity and binding targets is noisy, and traditional methods for analyzing such data may lead to inaccurate inferences, particularly with respect to combinatorial regulation. Hence, there is a need for robust and powerful approaches for better analyzing such data for accurate inferences.;In this dissertation, we develop statistical algorithms for analyzing transcription factor binding data. We model the uncertainty associated with such data, and study various aspects of combinatorial transcriptional regulation. We also develop an algorithm to quantify the effect of noise on various inferences from such data. Finally, we apply the Random Forests algorithm to identify DNA sequences that serve as the targets for transcription factors. |
