Assessing the Impacts of Small Variants in the Giant Sarcomeric Protein Titin Associated with Muscle Disease

Student thesis: Doctoral ThesisDoctor of Philosophy


Giant modular sarcomeric proteins such as titin have come under increasing scrutiny for their contributions to diseases such as skeletal and cardiac myopathies. Many variants in titin have been linked with these conditions through computational and experimental studies, and even relatively small, individual missense mutations have been shown to have far-reaching structural consequences. However, assessing the impact of all possible single nucleotide variants (SNVs) in titin is complicated by the immense size of this protein. By combining information on sequence, protein stability, and molecular dynamics features, I propose the generation of a variant classifier able to discriminate between neutral and deleterious variants. I review the disease-associated missense variants in titin reported in the literature and update the centralised resource for information relating to titin variants, TITINdb2, with newly available variants, annotations, structural models, and predictions. From here, I describe the current landscape of understanding of missense variants in titin I compare the coherence of predicted variant impacts with reported impacts in the literature, and investigate the potential of more computationally expensive features, using molecular dynamics, to add information lacking from static representations of these proteins. I found that various features associated with the movement of the protein domain backbone in simulation were strongly associated with the reported disease-associated missense variants in titin, but that these same features did not differentiate between strongly and weakly stability-altering variants. Furthermore, these results were consistent with machine learning models constructed to predict the pathogenicity or thermostability change associated with titin missense variants; features added to these classifiers from molecular dynamics and deep mutational scanning data improved their ability to discriminate between reported pathogenic and common population variants in titin over classifiers using sequence and structural data alone, but did not facilitate a reliable differentiation between different classes of thermostability-altering variants. These results reiterate the importance of greater diversity in information sources, including dynamical data, for the prediction of pathogenicity in missense variants, while also emphasising the future work required for the accurate determination of thermostability change, particularly for small missense variants.
Date of Award1 Aug 2023
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorFranca Fraternali (Supervisor) & Mathias Gautel (Supervisor)

Cite this