Interpretation of sequence variants in the biomedical environment: what should we take into account
2 likes1,084 views
The document discusses the prediction of pathogenicity in molecular mutations, highlighting the role of stability and function in assessing mutations like F367V. It emphasizes the lack of a direct relationship between stability and functional impact, and notes the approximations involved in current predictive modeling approaches like FoldX. Recommendations include using combined prediction strategies and careful analysis of mutations to improve accuracy in biomedical decisions.
Interpretation of sequence variants in the biomedical environment: what should we take into account
- 1. Xavier de la Cruz Translational Bioinformatics,VHIR
- 2. + Translational Bioinformatics Casandra Riera Natàlia Padilla Òscar Marín Elena Álvarez de la Campa ‘Amics del VHIR’
- 3. Context: where are we going? The prediction score: ◦ role ◦ components: stability, MSA Going beyond Take home points...
- 4. SC SC
- 5. Xue et al., Gene,cs in Medicine, 2014
- 6. Dixon-Salazar et al., Sci.Transl. Med., 2012 Identifying molecular damage
- 11. PATHOLOGICAL NEUTRAL SCORE ERROR MUTATION: F367V SC=1.3
- 12. What should we consider when we want to use pathogenicity predictions in a biomedical decision process Why in some cases we can successfully interpret the pathogenicity of sequence variants, and why we fail in others CRITICAL UNDERSTANDING OFTHE PRINCIPLES UNDERLYING THE PREDICTION OF PATHOLOGICAL MUTATIONS
- 13. F367V SC = ?????????????????????????????????????????????????? MUTATION MOLECULAR IMPACT
- 14. F367V SC = Change in function ≈ Change in stability MUTATION MOLECULAR IMPACT Yue et al., J Mol Biol, 2005
- 15. Fersht, Matouschek & Serrano J Mol Biol, 1992 FoldX: Guerois et al., J Mol Biol, 2002 APPROXIMATION: (i) eliminate terms (ii) fit to in vitro exper. data SC = SC(stability)
- 16. 0 50 100 150 200 250 300 350 400 450 500 -8 -6 -4 -2 0 2 4 RelativeActivity Stability (Kcal/mol) Bromber & Rost, BMC Bioinformatics, 2009 Phage T4 Lysozyme mutants r=0.3
- 17. There is not an exact relationship between function and stability SC(stability) is an approximation In vitro data are used to model stability in Foldx, etc
- 21. There is not an exact relationship between function and stability SC(stability) is an approximation In vitro data are used to model stability in Foldx, etc Structure gives a qualitative view of stability change Manual analysis of 3D structures does not provide quantitative information
- 22. F367V SC = SC(stability) + SC(function) MUTATION MOLECULAR IMPACT Yue et al., J Mol Biol, 2005
- 23. SC(function) = wAcS.AcS + wCons1.Cons1 + wCons2.Cons2 + ... Are we hitting the active site? (Yes/No) Do we know the active site? Usually No We can use residue conservation How to quantify it? Use MSA and some conservation measure
- 24. APPROXIMATIONS: (i) annotations may be imprecise (ii) what conservation measure (iii) building MSAs is computationally unsolved problem (NP-hard) (iv) the weighs are empirical SC(function) = wAcS.AcS + wCons1.Cons1 + wCons2.Cons2 + ...
- 25. Imprecise annotations (Yue et al., J Mol Biol, 2005) MSA are suboptimal Several conservation measures
- 26. STABILITY: ◦ There is not an exact relationship between function and stability ◦ SC(stability) is an approximation ◦ In vitro data are used to model stability in Foldx, etc ◦ Structure gives a qualitative view of stability change ◦ Manual analysis of 3D structures does not provide quantitative information FUNCTION: ◦ SC(function) is an approximation ◦ functional information is indirectly obtained from MSA: prediction depends on MSA (and these are an approximation)
- 27. Protein stability Functional interactions Unspecific cellular interactions …KKRHCSGWL… Y Change in function
- 28. SC = SC(stability) + SC(function) SC = SC(stability) + SC(function) MUTATION MOLECULAR IMPACT PATHOLOGICAL NEUTRAL
- 29. Combine several strategies: ◦ Only use the best predictions, however,... ◦ Use several methods ◦ Manual analysis (better through collaborations)
- 30. PATHOLOGICAL NEUTRAL NO PREDICTION • BETTER ACCURACY • LESS PREDICTIONS
- 31. Consistency at low scores may serve to indicate mild pathological mutations
- 32. STABILITY: ◦ There is not an exact relationship between function and stability ◦ SC(stability) is an approximation ◦ In vitro data are used to model stability in Foldx, etc ◦ Structure gives a qualitative view of stability change ◦ Manual analysis of 3D structures does not provide quantitative information FUNCTION: ◦ SC(function) is an approximation ◦ functional information is indirectly obtained from MSA: prediction depends on MSA (and these are an approximation) GOING BEYOND: ◦ take only the best predictions ◦ use several methods ◦ look for unexpected effects