The article reviews the recent advances in understanding and identifying metamorphic proteins, which can reversibly interconvert between multiple native states. Metamorphic proteins challenge the traditional notion that most proteins have a single, unique native state. The review highlights computational and experimental strategies for studying their refolding processes and identifying new metamorphic proteins. Key methods include molecular dynamics simulations, particularly dual-basin structure-based models (SBMs), which have been used to simulate the refolding pathways of various fold-switching proteins. Experimental techniques such as NMR spectroscopy and hydrogen-deuterium exchange mass spectrometry (HDX-MS) are also discussed for characterizing structural changes and timescales of fold switching. The article emphasizes the importance of evolutionary information in predicting metamorphic proteins, noting that coevolutionary analysis can identify unique contacts corresponding to alternative folds. However, it also discusses the limitations of artificial intelligence (AI)-based methods in predicting metamorphic proteins, highlighting the need for more robust training sets and better understanding of protein folding physics. Finally, the article outlines experimental assays for screening new fold-switching proteins, emphasizing the importance of these methods in advancing the field.The article reviews the recent advances in understanding and identifying metamorphic proteins, which can reversibly interconvert between multiple native states. Metamorphic proteins challenge the traditional notion that most proteins have a single, unique native state. The review highlights computational and experimental strategies for studying their refolding processes and identifying new metamorphic proteins. Key methods include molecular dynamics simulations, particularly dual-basin structure-based models (SBMs), which have been used to simulate the refolding pathways of various fold-switching proteins. Experimental techniques such as NMR spectroscopy and hydrogen-deuterium exchange mass spectrometry (HDX-MS) are also discussed for characterizing structural changes and timescales of fold switching. The article emphasizes the importance of evolutionary information in predicting metamorphic proteins, noting that coevolutionary analysis can identify unique contacts corresponding to alternative folds. However, it also discusses the limitations of artificial intelligence (AI)-based methods in predicting metamorphic proteins, highlighting the need for more robust training sets and better understanding of protein folding physics. Finally, the article outlines experimental assays for screening new fold-switching proteins, emphasizing the importance of these methods in advancing the field.