Limits...
Flanking p10 contribution and sequence bias in matrix based epitope prediction: revisiting the assumption of independent binding pockets.

Parry CS - BMC Struct. Biol. (2008)

Bottom Line: One new matrix shows significant improvement over the base matrix; the other does not.One of the extended quantitative matrices showed significant improvement in prediction over the original nine residue matrix and over the other extended matrix.Proline in the sequence of the peptide library of the better performing matrix presumably stabilizes the peptide conformation through neighbour interactions.

View Article: PubMed Central - HTML - PubMed

Affiliation: Computational Biophysics Section, Laboratory of Computational Biology, National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892-9314, USA. csparry@helix.nih.gov

ABSTRACT

Background: Eluted natural peptides from major histocompatibility molecules show patterns of conserved residues. Crystallographic structures show that the bound peptide in class II major histocompatibility complex adopts a near uniform polyproline II-like conformation. This way allele-specific favoured residues are able to anchor into pockets in the binding groove leaving other peptide side chains exposed for recognition by T cells. The anchor residues form a motif. This sequence pattern can be used to screen large sequences for potential epitopes. Quantitative matrices extend the motif idea to include the contribution of non-anchor peptide residues. This report examines two new matrices that extend the binding register to incorporate the polymorphic p10 pocket of human leukocyte antigen DR1. Their performance is quantified against experimental binding measurements and against the canonical nine-residue register matrix.

Results: One new matrix shows significant improvement over the base matrix; the other does not. The new matrices differ in the sequence of the peptide library.

Conclusion: One of the extended quantitative matrices showed significant improvement in prediction over the original nine residue matrix and over the other extended matrix. Proline in the sequence of the peptide library of the better performing matrix presumably stabilizes the peptide conformation through neighbour interactions. Such interactions may influence epitope prediction in this test of quantitative matrices. This calls into question the assumption of the independent contribution of individual binding pockets.

Show MeSH

Related in: MedlinePlus

Plots of binding measurements versus prediction values. Half maximal inhibitory concentration (IC50) values of peptide sequences are plotted as a function of their predicted values for each of the three matrices, P9 (open circles, blue), P10 (cross, red) and PP10 (stars, green). A line is fitted to the plotted values in the respective colors. This is done for data sets A. Glutamic acid decarboxylase, GAD65; B. Islet cell antigen protein, ICA69; and C. Varicella zoster virus, VZV.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2600787&req=5

Figure 1: Plots of binding measurements versus prediction values. Half maximal inhibitory concentration (IC50) values of peptide sequences are plotted as a function of their predicted values for each of the three matrices, P9 (open circles, blue), P10 (cross, red) and PP10 (stars, green). A line is fitted to the plotted values in the respective colors. This is done for data sets A. Glutamic acid decarboxylase, GAD65; B. Islet cell antigen protein, ICA69; and C. Varicella zoster virus, VZV.

Mentions: The best fit was found to be a straight line. Higher order polynomial functions were tried but gave worse results. Low IC50 values correspond to good binding and high IC50 values correspond to poor binding, and negative predicted values. A flat fit has no predictive use. Plots for GAD65, ICA69 and VZV are shown in Figures 1 and 2. In all three scoring methods, there are both false positives (upper right quadrant) and false negatives (bottom left quadrant, Figures 1 and 2). False positives can be screened out in validation tests but false negatives are problematic. With these data sets, the scoring matrices show few false negatives. This is a useful property and the matrices can be used to screen large sequences without missing potential epitopes. A high threshold may be set to eliminate falsely predicted peptides.


Flanking p10 contribution and sequence bias in matrix based epitope prediction: revisiting the assumption of independent binding pockets.

Parry CS - BMC Struct. Biol. (2008)

Plots of binding measurements versus prediction values. Half maximal inhibitory concentration (IC50) values of peptide sequences are plotted as a function of their predicted values for each of the three matrices, P9 (open circles, blue), P10 (cross, red) and PP10 (stars, green). A line is fitted to the plotted values in the respective colors. This is done for data sets A. Glutamic acid decarboxylase, GAD65; B. Islet cell antigen protein, ICA69; and C. Varicella zoster virus, VZV.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2600787&req=5

Figure 1: Plots of binding measurements versus prediction values. Half maximal inhibitory concentration (IC50) values of peptide sequences are plotted as a function of their predicted values for each of the three matrices, P9 (open circles, blue), P10 (cross, red) and PP10 (stars, green). A line is fitted to the plotted values in the respective colors. This is done for data sets A. Glutamic acid decarboxylase, GAD65; B. Islet cell antigen protein, ICA69; and C. Varicella zoster virus, VZV.
Mentions: The best fit was found to be a straight line. Higher order polynomial functions were tried but gave worse results. Low IC50 values correspond to good binding and high IC50 values correspond to poor binding, and negative predicted values. A flat fit has no predictive use. Plots for GAD65, ICA69 and VZV are shown in Figures 1 and 2. In all three scoring methods, there are both false positives (upper right quadrant) and false negatives (bottom left quadrant, Figures 1 and 2). False positives can be screened out in validation tests but false negatives are problematic. With these data sets, the scoring matrices show few false negatives. This is a useful property and the matrices can be used to screen large sequences without missing potential epitopes. A high threshold may be set to eliminate falsely predicted peptides.

Bottom Line: One new matrix shows significant improvement over the base matrix; the other does not.One of the extended quantitative matrices showed significant improvement in prediction over the original nine residue matrix and over the other extended matrix.Proline in the sequence of the peptide library of the better performing matrix presumably stabilizes the peptide conformation through neighbour interactions.

View Article: PubMed Central - HTML - PubMed

Affiliation: Computational Biophysics Section, Laboratory of Computational Biology, National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892-9314, USA. csparry@helix.nih.gov

ABSTRACT

Background: Eluted natural peptides from major histocompatibility molecules show patterns of conserved residues. Crystallographic structures show that the bound peptide in class II major histocompatibility complex adopts a near uniform polyproline II-like conformation. This way allele-specific favoured residues are able to anchor into pockets in the binding groove leaving other peptide side chains exposed for recognition by T cells. The anchor residues form a motif. This sequence pattern can be used to screen large sequences for potential epitopes. Quantitative matrices extend the motif idea to include the contribution of non-anchor peptide residues. This report examines two new matrices that extend the binding register to incorporate the polymorphic p10 pocket of human leukocyte antigen DR1. Their performance is quantified against experimental binding measurements and against the canonical nine-residue register matrix.

Results: One new matrix shows significant improvement over the base matrix; the other does not. The new matrices differ in the sequence of the peptide library.

Conclusion: One of the extended quantitative matrices showed significant improvement in prediction over the original nine residue matrix and over the other extended matrix. Proline in the sequence of the peptide library of the better performing matrix presumably stabilizes the peptide conformation through neighbour interactions. Such interactions may influence epitope prediction in this test of quantitative matrices. This calls into question the assumption of the independent contribution of individual binding pockets.

Show MeSH
Related in: MedlinePlus