Molecular probes are often at the forefront of chemical attempts to investigate biological structures for their function. Indeed, molecular probes that bind the G-quadruplex (G4) nucleic acid structure have been instrumental in demonstrating the existence and biological importance of these non-canonical secondary structures. There is increasing evidence that G4s present new opportunities in treating cancer and neurodegenerative disease, but a deeper understanding of the mechanistic logic behind G4 function is necessary to exploit this secondary structure as a clinical target. The generation of new, high-quality G4 probes can address this challenge. Firstly, investigating G4s with molecular probes becomes especially powerful when chemically contrasting G4 ligands reveal similar observations in a biological context. Secondly, molecular probes often spearhead the development of new methods of therapeutic intervention. This thesis commences the exploration of cyclic peptides as alternative chemical scaffolds for de novo G4 ligand generation. Widely used to generate protein binders as drug leads but relatively under-explored for nucleic acid targeting, cyclic peptides lie in the chemical space between small molecules and antibodies and combine advantages of both. In particular, the rich 3D stereochemical complexity of cyclic peptides allows these molecules to access binding modes resembling that of a protein active site, and yet retain a small molecule’s ease of chemical manipulation. Moreover, since amino acids are genetically encodable, they naturally lend themselves to biosynthetic, high-throughput selection methodologies, allowing simultaneous analysis and identification from libraries containing millions of different molecules. Two contrasting cyclic peptide systems are focused on; the first generates bicyclic peptides (so-called ‘bicycles’) with phage display. Cysteine-rich peptides displayed on the virus coats are readily modified in situ to generate libraries (10$9$) of bicyclic peptides. These libraries are sequentially screened against immobilised targets to artificially select ligands with the strongest biophysical interaction. With optimisation of the phage display selection process and motif analysis, a bicyclic peptide lead (ckit1bpep4) was generated with G4 binding properties comparable to those of the most utilised small molecule ligands. Additionally, it was demonstrated that ckit1bpep4 could be readily modified for usage as a chemical tool without compromising these binding properties and subsequently imaged in cells. SICLOPPS, the second methodology, allows genetic expression of head-tail cyclised peptides in living systems. SICLOPPS libraries allow generation and selection of ligands that lead to a biological response, as opposed to a purely biophysical interaction. In this way, the genetic selection evaluates effects such as cellular bioavaliability and off-target effects that conventional display techniques cannot access. After porting the SICLOPPS system for lentiviral transduction, we carried out such a genetic selection in DT40 cells, assessing expression loss of a surface protein BU-1 produced by an endogenous G4-containing locus. Peptides identified that disrupt BU-1 expression were then validated against a cell line with the BU-1 G4 edited out, allowing validation of G4 specificity. These cyclic peptides were then synthesised and biophysically validated to bind G4s. Overall, this thesis demonstrates the rich potential of cyclic peptides as G4 ligands. Their favourable binding properties and powerful generation methodologies suggest that cyclic peptides could form a platform for producing ligands for nucleic acids as well as proteins. Such probes have the potential to advance our knowledge of nucleic acids biology and inspire novel therapeutics.