Applied Computational and Systems Biology – examquestions
Part 1. Structural computational biology - Schymkowitz
3 to 4 open questions: 3 questions assess knowledge of the subject; one question asks for insight.
Why are proteins domains limited in size? What is the evolutionary solution to this problem? (CH1, p.35)
• Proteins are limited in domain size due to the constraint that hydrophobic stretches become too long,
making them insoluble . The evolutionary solution to this problem is the existence of multidomain
proteins, where different functional units (domains) are combined into a single protein molecule . This
arrangement allows for the incorporation of diverse functionalities within a single protein structure,
enabling greater complexity and versatility in biological processes
What is the role of the electrostatic interaction in protein stability and protein-protein interactions? (CH1,
p.41; p.47; p.51)
• Electrostatic interactions play a crucial role in protein stability and protein-protein interactions by
contributing to the overall energetics and dynamics of these biological processes.
• 1. Protein Stability:
- Electrostatic interactions between charged amino acid residues within a protein can stabilize its
folded structure through salt bridges and ion pairs. These interactions help in maintaining the
specific conformation of the protein by contributing to the overall stability of the folded state [T5].
- The formation of salt bridges between oppositely charged residues can provide significant
contributions to the enthalpic component of protein stability, thereby influencing the overall folding
and structural integrity of the protein [T6].
• 2. Protein-Protein Interactions:
- Electrostatic interactions between proteins play a vital role in mediating protein-protein
recognition and binding. Oppositely charged regions on different proteins can attract each other,
leading to the formation of stable complexes.
- In protein-protein interactions, electrostatic forces can contribute to the overall binding affinity and
specificity between interacting proteins. The long-range nature of electrostatic interactions allows
proteins to sense each other from a distance, facilitating the initial recognition and association
process [T6].
- The electrostatic complementarity between interacting proteins can influence the binding kinetics
and thermodynamics of the interaction. Electrostatic interactions can contribute to both the
enthalpic and entropic components of the binding free energy, affecting the overall stability of the
protein-protein complex [T6].
• Overall, electrostatic interactions play a significant role in protein stability by contributing to the folding
and structural integrity of proteins through salt bridges and ion pairs. In protein-protein interactions,
electrostatic forces mediate the recognition, binding, and stability of protein complexes by influencing
the binding affinity, specificity, and dynamics of the interaction [T5], [T6].
,(half a page): Explain the different 'quality metrics' used for protein structures. Explain what they mean and
how they can be different for experimental structures and predicted structures. (CH0, p.5), CH2
Give short the structure determination methods and compare (CH2, p.27)
• Structure determination methods in structural computational biology involve various techniques for
elucidating the three-dimensional structures of biological macromolecules such as proteins and
nucleic acids. Here is a brief overview and comparison of some common structure determination
methods:
• 1. X-ray Crystallography:
- Principle: X-ray diffraction patterns from crystallized samples are used to determine the electron
density map and infer the atomic structure.
- Strengths: High resolution, suitable for a wide range of molecular sizes, and provides a static
snapshot of the structure.
- Weaknesses: Requires crystallization, which can be time-consuming and challenging for some
proteins. May yield artifacts in the structure [T5].
• 2. Nuclear Magnetic Resonance (NMR) Spectroscopy:
- Principle: Measures interactions between nuclear spins in a magnetic field to determine the
atomic-level structure and dynamics of molecules in solution.
- Strengths: Provides dynamic information, suitable for studying proteins in solution, and can
determine structures of smaller proteins.
- Weaknesses: Limited to proteins below a certain size limit (50-70 kDa), typically lower resolution
compared to X-ray crystallography [T5].
• 3. Cryo-Electron Microscopy (Cryo-EM):
- Principle: Utilizes electron microscopy of samples maintained at cryogenic temperatures to
visualize macromolecular structures.
- Strengths: Allows direct observation of large macromolecular complexes, does not require
crystallization, and provides some ensemble information.
- Weaknesses: Requires high molecular mass (lower limit around 50 kDa), and the resolution may
vary depending on the sample and imaging conditions [T5].
• Comparison:
- X-ray crystallography offers high resolution but requires crystallization, while NMR provides
dynamic information but is limited by the size of the protein.
- Cryo-EM allows the study of large complexes without crystallization but may have resolution
limitations compared to X-ray crystallography and NMR.
- Each method has its strengths and weaknesses, making them complementary in providing
structural insights into biological macromolecules.
- The choice of structure determination method depends on the size, nature, and properties of the
molecule being studied, as well as the research objectives and available resources.
• By utilizing a combination of these structure determination methods, researchers can obtain
comprehensive structural information to understand the biological functions and mechanisms of
macromolecules.
, Prediction metrices (CH0, p.5), (CH3, p.6)
• Prediction metrics are essential tools used to evaluate the performance of computational models and
algorithms in predicting various outcomes, such as the impact of mutations on protein function or the
accuracy of structural predictions. Here are some common prediction metrics used in computational
biology and bioinformatics:
• 1. Accuracy:
- Definition: The proportion of correct predictions made by the model.
- Formula: (TP + TN) / (TP + FP + FN + TN)
- Interpretation: Provides an overall measure of the model's correctness in predicting both positive
and negative outcomes
• 2. Precision:
- Definition: The proportion of predicted positives that are truly positive.
- Formula: TP / (TP + FP)
- Interpretation: Indicates the accuracy of positive predictions made by the model, minimizing false
positives.
• 3. Recall (Sensitivity):
- Definition: The fraction of true positives correctly predicted by the model.
- Formula: TP / (TP + FN)
- Interpretation: Measures the model's ability to identify all relevant instances, minimizing false
negatives.
• 4. False Positive Rate:
- Definition: The fraction of false positive predictions made by the model.
- Formula: FP / (TN + FP)
- Interpretation: Evaluates the rate of incorrect positive predictions relative to the actual negative
instances.
• 5. Matthews Correlation Coefficient (MCC):
- Definition: A correlation coefficient that takes into account true positives, true negatives, false
positives, and false negatives.
- Formula: (TP * TN - FP * FN) / sqrt((TP + FP) * (TP + FN) * (TN + FP) * (TN + FN))
- Interpretation: Provides a balanced measure of the model's performance, particularly useful for
imbalanced datasets.
• 6. Area Under the Curve (AUC):
- Definition: A metric that evaluates the performance of binary classification models across different
thresholds.
- Interpretation: AUC values range from 0 to 1, where 0.5 indicates random predictions, 0 represents
anti-correlation, and 1 signifies perfect correlation between predicted and actual outcomes.
• These prediction metrics are commonly used to assess the accuracy, precision, recall, and overall
performance of computational models in various bioinformatics applications, including protein
structure prediction, mutation impact analysis, and functional annotation. Researchers use these
metrics to evaluate and compare the effectiveness of different algorithms and approaches in predicting
biological outcomes.
