Background ⎯ Malignant melanoma is an aggressive form of skin cancer with poor prognosis. Current immunotherapies targeting melanoma using the patient’s immune system can achieve long-­‐term melanoma clearance in some patients. T-­‐cells are the main effectors of this anti-­‐melanoma immunity and can specifically recognise and lyse tumour cells. T-­‐cells express unique surface T-­‐cell receptors (TCR) which recognise tumour-­‐derived peptide antigens bound to a Human Leukocyte Antigen (HLA) molecule on melanoma cells. My research aimed at studying anti-­‐melanoma T-­‐cell responses from both a molecular and cellular level. Results ⎯ Initially, I analysed the first crystallographic structure of a TCR in complex with a peptide from the melanoma protein Glycoprotein(gp)100 bound to HLA-­‐A2. I show that an unanticipated molecular switch of the position(P)4-­‐5 peptide bond by substitution of the 3 residue could brogate recognition by host TCRs. I then dissected successful T-­‐cell responses, restricted to specific HLA alleles, in melanoma infiltrates from two complete responder (CR) patients after Tumour Infiltrating Lymphocyte (TIL) therapy. The antigen specificity and phenotype of anti-­‐melanoma TILs were evaluated, and a panel of TIL clones capable of lysing autologous melanoma cell targets was characterised. Interestingly, in each patient studied, melanoma-­‐specific clonotypes detected Background⎯ Malignant melanoma is an aggressive form of skin cancer with poor prognosis. Current immunotherapies targeting melanoma using the patient’s immune system can achieve long-­‐term melanoma clearance in some patients. T-­‐cells are the main effectors of this anti-­‐melanoma immunity and can specifically recognise and lyse tumour cells. T-­‐cells express unique surface T-­‐cell receptors (TCR) which recognise tumour-­‐derived peptide antigens bound to a Human Leukocyte Antigen (HLA) molecule on melanoma cells. My research aimed at studying anti-­‐melanoma T-­‐cell responses from both a molecular and cellular level. Results ⎯ Initially, I analysed the first crystallographic structure of a TCR in complex with a peptide from the melanoma protein Glycoprotein(gp)100 bound to HLA-­‐A2. I show that an unanticipated molecular switch of the position(P)4-­‐5 peptide bond by substitution of the P3 residue could abrogate recognition by host TCRs. I then dissected successful T-­‐cell responses, restricted to specific HLA alleles, in melanomainfiltrates from two complete responder (CR) patients after Tumour Infiltrating Lymphocyte (TIL) therapy. The antigen specificity and phenotype of anti-­‐melanoma TILs were evaluated, and a panel of TIL clones capable of lysing autologous melanoma cell targets was characterised. Interestingly, in each patient studied, melanoma-­‐specific clonotypes detected in the TILs persist in the blood after TIL therapy and recognise T-­‐cell epitopes shared by other patients and other tumour types. Conclusions ⎯ I demonstrate that structural studies should be considered when designing improved peptide cancer vaccines as even single substitutions in residues not heavily engaged by the TCR can have important, unpredictable knock-­‐on effects that impair T-­‐cell recognition. Understanding the key antigen-­‐specificities of anti-­‐tumour TILs from CR patients, will improve the efficacy of T-­‐cell based therapies. Validated antigens could be applied as biomarkers of, or targets for, cancer immunotherapy. Dominant shared antigens targeted in CR patients may make promising candidates for therapeutic vaccination.