Project Summary

The study focuses on refining predictions of transcription factor (TF) binding sites through convolutional neural networks. It aims to leverage TF ChIP-seq, ATAC-seq, and histone modification peaks data for enhanced predictive accuracy, exploring optimal architectures and data representations.

Employing data processing methods, we encoded sequence data into a matrix, utilizing a model that combines Conv1D, LSTM, and dense layers. The approach resulted in a model that shows promising accuracy in its predictions.

Future directions include efforts to improve model precision and the potential to generalize the model for broader applications, highlighting the ongoing quest for more effective computational tools in genomic research.