Transcription factors (TF) are key proteins involved in gene regulation by binding to specific DNA sites. The C₂H₂ zinc finger (ZF) TFs form the largest family of DNA binding proteins in eukaryotes and are a key participant in the regulation of most genes. A major obstacle towards understanding the molecular basis of transcriptional regulation is the lack of a general recognition code for protein-DNA interactions. In this thesis, we aim to understand molecular mechanisms of DNA binding/recognition by TFs and to quantitatively estimate recognition rules for TF-DNA interactions. To this aim, we first identified key residues that play an important role in ZF-DNA binding and studied their dynamics prior to binding using molecular dynamics (MD) simulations. We found that key residues that are buried upon complexation are prealigned to conformations close to their bound state prior to binding. The bound-like behavior of some of these residues is found to be dependent on the ion concentration of the system, consistent with experimental observations of increased binding affinity with increased ionic strength in protein-DNA interactions. We identified a binding site for Cl^- ions located in the same pocket where DNA phosphates are found most buried in the complex structure of ZFs. Bound Cl^- ions constrain key side chains in conformations similar to those observed when interacting with the phosphates. These results suggest that ZFs are able to maintain bound like conformations of key residues upon encountering the DNA hinting at a general mechanism to rapidly form encounter complexes amenable for a fast readout of the DNA. Next, we develop a novel experimentally-based approach using high quality crystal structures and binding data on the promiscuous family of C₂H₂ zinc fingers (ZF) and decode ten fundamental specific interactions responsible for protein-DNA recognition. The interactions include five hydrogen bond types, three atomic desolvation penalties, a favorable non-polar energy, and a novel water accessibility factor. We apply this code to three large data sets containing a total of 89 C₂H₂ TF mutants on the three ZFs of EGR. Guided by MD simulations of individual ZFs, we map the interactions into homology models that embody all feasible intra- and inter-molecular bonds, selecting for each sequence the structure with the lowest free energy. The interactions reproduce the change in affinity of 35 mutants of finger I (R² = 0.99), as well as two independent validation sets of 23 mutants of finger II (R ² = 0.97) and 31 human ZFs on finger IIII (R² = 0.95). More importantly, the method predicts bound ZF-DNA complexes for all 89 mutants, decoding molecular basis of EGR-DNA specificity. Our findings reveal recognition rules that depend on DNA sequence/structure, molecular water at the interface and induced fit of the C₂H₂ TFs. Collectively, our method provides the first robust framework to decode the molecular basis of TFs binding to DNA.