-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathalbanese_dissertation.toc
125 lines (125 loc) · 12.1 KB
/
albanese_dissertation.toc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
\gobblefive
\contentsline {section}{Dedication}{iii}{Doc-Start}
\gobblefive
\contentsline {section}{Biographical Sketch}{vi}{Doc-Start}
\gobblefive
\contentsline {section}{Acknowledgements}{vii}{chapter*.1}
\gobblefive
\contentsline {section}{Table of Contents}{x}{chapter*.2}
\contentsline {section}{List of Tables}{xiii}{chapter*.2}
\contentsline {section}{List of Figures}{xiv}{chapter*.2}
\contentsline {chapter}{\numberline {1}Introduction}{1}{chapter.1}
\contentsline {section}{\numberline {1.1}Perspective}{1}{section.1.1}
\contentsline {section}{\numberline {1.2}Synopsis}{5}{section.1.2}
\contentsline {chapter}{\numberline {2}Predicting the selectivity of small molecule kinase inhibitors}{7}{chapter.2}
\contentsline {section}{\numberline {2.1}Gloss}{7}{section.2.1}
\contentsline {section}{\numberline {2.2}Abstract}{10}{section.2.2}
\contentsline {section}{\numberline {2.3}Introduction}{11}{section.2.3}
\contentsline {subsection}{\numberline {2.3.1}Free energy methods can aid structure-based drug design}{11}{subsection.2.3.1}
\contentsline {subsection}{\numberline {2.3.2}Selectivity is an important consideration in drug design}{11}{subsection.2.3.2}
\contentsline {subsection}{\numberline {2.3.3}The use of physical modeling to predict selectivity is relatively unexplored}{12}{subsection.2.3.3}
\contentsline {subsection}{\numberline {2.3.4}Kinases are an important and particularly challenging model system for selectivity predictions}{13}{subsection.2.3.4}
\contentsline {subsection}{\numberline {2.3.5}The correlation coefficient measures how useful predictions are in achieving selectivity}{15}{subsection.2.3.5}
\contentsline {section}{\numberline {2.4}Results}{17}{section.2.4}
\contentsline {subsection}{\numberline {2.4.1}Alchemical free energy methods can be used to predict compound selectivity}{17}{subsection.2.4.1}
\contentsline {subsection}{\numberline {2.4.2}Correlation in force field errors can significantly enhance accuracy of selectivity predictions}{18}{subsection.2.4.2}
\contentsline {subsection}{\numberline {2.4.3}Poor selectivity is achieved for the closely related kinases CDK2/CDK9}{22}{subsection.2.4.3}
\contentsline {subsection}{\numberline {2.4.4}Greater selectivity is achieved for more distantly related kinases CDK2/ERK2}{25}{subsection.2.4.4}
\contentsline {subsection}{\numberline {2.4.5}FEP+ calculations show smaller than expected errors for $\Delta S$ predictions}{28}{subsection.2.4.5}
\contentsline {subsection}{\numberline {2.4.6}Correlation of forcefield errors accelerates selectivity optimization}{31}{subsection.2.4.6}
\contentsline {subsection}{\numberline {2.4.7}Expending more effort to reduce statistical error can be beneficial in selectivity optimization}{34}{subsection.2.4.7}
\contentsline {section}{\numberline {2.5}Discussion and Conclusions}{37}{section.2.5}
\contentsline {paragraph}{$S$ is a useful metric for selectivity in lead optimization}{37}{section*.9}
\contentsline {paragraph}{Forcefield error correlation can accelerate selectivity optimization}{37}{section*.10}
\contentsline {paragraph}{Two pairs of kinase test systems suggest forcefield errors can be correlated}{38}{section*.11}
\contentsline {paragraph}{Reducing statistical error is beneficial when forcefield errors are correlated}{39}{section*.12}
\contentsline {paragraph}{Larger data sets with a wide range of protein targets will enable future work}{40}{section*.13}
\contentsline {section}{\numberline {2.6}Methods}{41}{section.2.6}
\contentsline {subsection}{\numberline {2.6.1}Numerical model of selectivity optimization speedup}{41}{subsection.2.6.1}
\contentsline {subsection}{\numberline {2.6.2}Numerical model of impact of statistical error on selectivity optimization}{42}{subsection.2.6.2}
\contentsline {subsection}{\numberline {2.6.3}Binding Site Similarity analysis}{43}{subsection.2.6.3}
\contentsline {subsection}{\numberline {2.6.4}Extracting the binding free energy $\Delta G$ from reported experimental data}{44}{subsection.2.6.4}
\contentsline {subsection}{\numberline {2.6.5}Structure Preparation}{45}{subsection.2.6.5}
\contentsline {paragraph}{Ligand Pose Generation}{46}{section*.14}
\contentsline {subsection}{\numberline {2.6.6}Free Energy Calculations}{46}{subsection.2.6.6}
\contentsline {subsection}{\numberline {2.6.7}Statistical Analysis of FEP+ calculations}{47}{subsection.2.6.7}
\contentsline {subsection}{\numberline {2.6.8}Quantification of the correlation coefficient $\rho $}{48}{subsection.2.6.8}
\contentsline {subsection}{\numberline {2.6.9}Calculating the marginal distribution of speedup}{50}{subsection.2.6.9}
\contentsline {section}{\numberline {2.7}Acknowledgments}{50}{section.2.7}
\contentsline {section}{\numberline {2.8}Funding}{51}{section.2.8}
\contentsline {section}{\numberline {2.9}Disclosures}{51}{section.2.9}
\contentsline {section}{\numberline {2.10}Author Contributions}{52}{section.2.10}
\contentsline {chapter}{\numberline {3}Predicting the impact of clinically-observed kinase mutations using physical modeling}{53}{chapter.3}
\contentsline {section}{\numberline {3.1}Gloss}{53}{section.3.1}
\contentsline {section}{\numberline {3.2}Abstract}{57}{section.3.2}
\contentsline {section}{\numberline {3.3}Introduction}{57}{section.3.3}
\contentsline {subsection}{\numberline {3.3.1}The long tail of rare kinase mutations frustrates prediction of drug resistance}{58}{subsection.3.3.1}
\contentsline {subsection}{\numberline {3.3.2}Alchemical free-energy methods can predict inhibitor binding affinities}{61}{subsection.3.3.2}
\contentsline {subsection}{\numberline {3.3.3}Alchemical approaches can predict the impact of protein mutations on free energy}{61}{subsection.3.3.3}
\contentsline {subsection}{\numberline {3.3.4}Assessing the potential for physical modeling to predict resistance to FDA-approved TKIs}{62}{subsection.3.3.4}
\contentsline {section}{\numberline {3.4}Results}{63}{section.3.4}
\contentsline {subsection}{\numberline {3.4.1}A benchmark of $\Delta $pIC$_{50}$s for predicting mutational resistance}{63}{subsection.3.4.1}
\contentsline {subsection}{\numberline {3.4.2}Most mutations do not significantly reduce TKI potency}{69}{subsection.3.4.2}
\contentsline {subsection}{\numberline {3.4.3}FEP+ predicts affinity changes for clinical Abl mutants}{72}{subsection.3.4.3}
\contentsline {subsection}{\numberline {3.4.4}FEP+ accurately classifies affinity changes for Abl mutants}{76}{subsection.3.4.4}
\contentsline {subsection}{\numberline {3.4.5}How reliant are classification results on choice of cutoff?}{77}{subsection.3.4.5}
\contentsline {subsection}{\numberline {3.4.6}Bayesian analysis can estimate the true error}{79}{subsection.3.4.6}
\contentsline {subsection}{\numberline {3.4.7}How transferable is FEP+ across the six TKIs?}{80}{subsection.3.4.7}
\contentsline {subsection}{\numberline {3.4.8}Understanding the origin of mispredictions}{83}{subsection.3.4.8}
\contentsline {subsection}{\numberline {3.4.9}How strongly is accuracy affected for docked TKIs?}{84}{subsection.3.4.9}
\contentsline {section}{\numberline {3.5}Discussion and Conclusions}{88}{section.3.5}
\contentsline {paragraph}{Physics-based modeling can reliably predict when a mutation elicits resistance to therapy}{88}{section*.25}
\contentsline {paragraph}{Hierarchical Bayesian model estimates global performance}{91}{section*.27}
\contentsline {paragraph}{Experimentally observed IC$_{50}$ changes can be caused by other physical mechanisms}{92}{section*.28}
\contentsline {paragraph}{Other physical mechanisms of resistance are likely similarly computable.}{93}{section*.29}
\contentsline {paragraph}{Conclusion}{93}{section*.30}
\contentsline {section}{\numberline {3.6}Methods}{95}{section.3.6}
\contentsline {subsection}{\numberline {3.6.1}System preparation}{95}{subsection.3.6.1}
\contentsline {subsubsection}{Complexes with co-crystal structures.}{95}{section*.31}
\contentsline {subsubsection}{Complexes without co-crystal structures.}{99}{section*.33}
\contentsline {subsection}{\numberline {3.6.2}Force field parameter assignment}{100}{subsection.3.6.2}
\contentsline {subsection}{\numberline {3.6.3}Prime (MM-GBSA)}{100}{subsection.3.6.3}
\contentsline {subsection}{\numberline {3.6.4}Alchemical free energy perturbation calculations using FEP+}{101}{subsection.3.6.4}
\contentsline {subsection}{\numberline {3.6.5}Obtaining $\Delta \Delta $G from $\Delta $pIC$_{50}$ benchmark set data}{103}{subsection.3.6.5}
\contentsline {subsection}{\numberline {3.6.6}Assessing prediction performance}{105}{subsection.3.6.6}
\contentsline {subsubsection}{Quantitative accuracy metrics}{105}{section*.34}
\contentsline {subsubsection}{Truth tables}{105}{section*.35}
\contentsline {subsubsection}{Consensus model}{106}{section*.36}
\contentsline {subsubsection}{ROC}{106}{section*.37}
\contentsline {subsubsection}{Estimating uncertainties of physical-modeling results}{106}{section*.38}
\contentsline {subsubsection}{Bayesian hierarchical model to estimate intrinsic error}{107}{section*.39}
\contentsline {subsection}{\numberline {3.6.7}Data availability}{109}{subsection.3.6.7}
\contentsline {subsection}{\numberline {3.6.8}Code availability}{109}{subsection.3.6.8}
\contentsline {section}{\numberline {3.7}Acknowledgments}{110}{section.3.7}
\contentsline {section}{\numberline {3.8}Author Contributions}{110}{section.3.8}
\contentsline {section}{\numberline {3.9}Competing Interests}{110}{section.3.9}
\contentsline {chapter}{\numberline {4}Enabling high-throughput biophysical experiments on clinically-observed mutations}{111}{chapter.4}
\contentsline {section}{\numberline {4.1}Gloss}{111}{section.4.1}
\contentsline {section}{\numberline {4.2}Abstract}{114}{section.4.2}
\contentsline {section}{\numberline {4.3}Introduction}{115}{section.4.3}
\contentsline {section}{\numberline {4.4}Results}{119}{section.4.4}
\contentsline {subsection}{\numberline {4.4.1}Construct boundary choice impacts Abl kinase domain expression}{119}{subsection.4.4.1}
\contentsline {subsection}{\numberline {4.4.2}Screen of 96 kinases finds 52 with useful levels of automated \emph {E.\nobreakspace {}coli} expression}{123}{subsection.4.4.2}
\contentsline {subsection}{\numberline {4.4.3}High-expressing kinases are folded with a well-formed ATP binding site}{129}{subsection.4.4.3}
\contentsline {subsection}{\numberline {4.4.4}Fluorescence-based thermostability assay}{131}{subsection.4.4.4}
\contentsline {subsection}{\numberline {4.4.5}ATP-competitive inhibitor binding fluorescence assay}{132}{subsection.4.4.5}
\contentsline {subsection}{\numberline {4.4.6}Expressing clinically-derived Src and Abl mutants}{135}{subsection.4.4.6}
\contentsline {section}{\numberline {4.5}Discussion}{141}{section.4.5}
\contentsline {section}{\numberline {4.6}Methods}{143}{section.4.6}
\contentsline {subsection}{\numberline {4.6.1}Semi-automated selection of kinase construct sequences for \emph {E.\nobreakspace {}coli} expression}{143}{subsection.4.6.1}
\contentsline {subsubsection}{Selection of human protein kinase domain targets}{143}{section*.46}
\contentsline {subsubsection}{Matching target sequences with relevant PDB constructs}{144}{section*.47}
\contentsline {subsubsection}{Plasmid libraries}{145}{section*.48}
\contentsline {subsubsection}{Selection of sequence constructs for expression}{146}{section*.49}
\contentsline {subsubsection}{Automation of the construct selection process}{147}{section*.50}
\contentsline {subsection}{\numberline {4.6.2}Mutagenesis protocol}{147}{subsection.4.6.2}
\contentsline {subsection}{\numberline {4.6.3}Expression testing}{148}{subsection.4.6.3}
\contentsline {subsection}{\numberline {4.6.4}Fluorescence-based thermostability assay}{149}{subsection.4.6.4}
\contentsline {subsection}{\numberline {4.6.5}ATP-competitive inhibitor binding fluorescence assay}{151}{subsection.4.6.5}
\contentsline {subsection}{\numberline {4.6.6}Large Scale expression and purification protocol for MK14}{154}{subsection.4.6.6}
\contentsline {section}{\numberline {4.7}Author Contributions}{155}{section.4.7}
\contentsline {section}{\numberline {4.8}Acknowledgments}{156}{section.4.8}
\contentsline {chapter}{\numberline {5}Conclusion}{157}{chapter.5}
\contentsline {chapter}{Appendices}{160}{section*.51}
\contentsline {chapter}{\numberline {A}Supplemental Figures from Chapter 2}{161}{Appendix.1.A}
\contentsline {chapter}{Bibliography}{172}{table.caption.61}