main.toc

\babel@toc {english}{}\relax 
\gdef \the@ipfilectr {@-1}
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-2}
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-3}
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-4}
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-5}
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-6}
\contentsline {chapter}{\numberline {1}Introduction}{1}{chapter.1}%
\contentsline {section}{\numberline {1.1}Single cell analysis}{1}{section.1.1}%
\contentsline {subsection}{\numberline {1.1.1}Biological background}{1}{subsection.1.1.1}%
\contentsline {subsection}{\numberline {1.1.2}What do we measure and how?}{2}{subsection.1.1.2}%
\contentsline {subsection}{\numberline {1.1.3}High througput profiling methods}{3}{subsection.1.1.3}%
\contentsline {subsection}{\numberline {1.1.4}Data characteristics}{3}{subsection.1.1.4}%
\contentsline {subsection}{\numberline {1.1.5}Typical analysis workflow}{4}{subsection.1.1.5}%
\contentsline {subsubsection}{\nonumberline Exploratory data analysis}{4}{subsubsection*.13}%
\contentsline {subsubsection}{\nonumberline Feature selection and dimensionality reduction}{6}{subsubsection*.15}%
\contentsline {paragraph}{\nonumberline Highly variable gene selection}{7}{paragraph*.17}%
\contentsline {paragraph}{\nonumberline Dimensionality reduction}{7}{paragraph*.19}%
\contentsline {paragraph}{\nonumberline 2D cell-level vizualisation}{8}{paragraph*.21}%
\contentsline {subsubsection}{\nonumberline Clustering}{8}{subsubsection*.23}%
\contentsline {subsubsection}{\nonumberline Batch effect correction}{9}{subsubsection*.25}%
\contentsline {subsubsection}{\nonumberline Trajectory inference}{10}{subsubsection*.27}%
\contentsline {paragraph}{\nonumberline Cluster-based TI approaches}{10}{paragraph*.29}%
\contentsline {paragraph}{\nonumberline Dynamic time warping}{10}{paragraph*.31}%
\contentsline {paragraph}{\nonumberline Minimum weight bipartite matching}{11}{paragraph*.33}%
\contentsline {section}{\numberline {1.2}Challenges in single-cell omics}{12}{section.1.2}%
\contentsline {subsection}{\numberline {1.2.1}Trajectory inference and dynamical systems}{12}{subsection.1.2.1}%
\contentsline {subsection}{\numberline {1.2.2}R/Python interoperability}{13}{subsection.1.2.2}%
\contentsline {subsection}{\numberline {1.2.3}Standardizing benchmarks}{14}{subsection.1.2.3}%
\contentsline {section}{\numberline {1.3}Research context and objectives}{16}{section.1.3}%
\contentsline {paragraph}{\nonumberline Trajectory inference and dynamical systems}{16}{paragraph*.39}%
\contentsline {paragraph}{\nonumberline Interoperability}{17}{paragraph*.41}%
\contentsline {paragraph}{\nonumberline Benchmarking visualization}{17}{paragraph*.43}%
\gdef \the@ipfilectr {}
\contentsline {part}{\numberline {I}Trajectory inference}{25}{part.1}%
\gdef \the@ipfilectr {@-7}
\contentsline {chapter}{\numberline {2}Recent advances in trajectory inference from single-cell omics data}{27}{chapter.2}%
\contentsline {section}{\numberline {2.1}Introduction}{28}{section.2.1}%
\contentsline {section}{\numberline {2.2}Trajectory topologies}{30}{section.2.2}%
\contentsline {section}{\numberline {2.3}Computational approaches for trajectory inference}{33}{section.2.3}%
\contentsline {subsection}{\numberline {2.3.1}Current approaches}{33}{subsection.2.3.1}%
\contentsline {subsection}{\numberline {2.3.2}Novel probabilistic approaches}{34}{subsection.2.3.2}%
\contentsline {section}{\numberline {2.4}Extensions of trajectory inference}{36}{section.2.4}%
\contentsline {subsection}{\numberline {2.4.1}Integration of other data sources}{36}{subsection.2.4.1}%
\contentsline {subsection}{\numberline {2.4.2}Quality control and benchmarking}{37}{subsection.2.4.2}%
\contentsline {section}{\numberline {2.5}Downstream analysis of inferred trajectories}{37}{section.2.5}%
\contentsline {subsection}{\numberline {2.5.1}Visualisation}{38}{subsection.2.5.1}%
\contentsline {subsection}{\numberline {2.5.2}Trajectory differential expression}{38}{subsection.2.5.2}%
\contentsline {subsection}{\numberline {2.5.3}Trajectory alignment and comparison}{39}{subsection.2.5.3}%
\contentsline {subsection}{\numberline {2.5.4}Dynamic gene regulatory network inference}{39}{subsection.2.5.4}%
\contentsline {section}{\numberline {2.6}A future outlook for trajectory inference methods}{40}{section.2.6}%
\contentsline {section}{\numberline {2.7}Acknowledgements}{42}{section.2.7}%
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-8}
\contentsline {chapter}{\numberline {3}dynchro: exploratory analysis and alignment of inferred trajectories}{51}{chapter.3}%
\contentsline {section}{\numberline {3.1}Introduction}{52}{section.3.1}%
\contentsline {section}{\numberline {3.2}Results}{54}{section.3.2}%
\contentsline {subsection}{\numberline {3.2.1}dynchro workflow}{54}{subsection.3.2.1}%
\contentsline {subsection}{\numberline {3.2.2}dynchro shows the effect of batch removal on trajectories and identifies premature stops}{56}{subsection.3.2.2}%
\contentsline {subsection}{\numberline {3.2.3}dynchro identifies correspondences between branching trajectories}{59}{subsection.3.2.3}%
\contentsline {subsection}{\numberline {3.2.4}dynchro clusters similar and dissimilar samples based on trajectories}{61}{subsection.3.2.4}%
\contentsline {subsection}{\numberline {3.2.5}dynchro identifies and tracks similarity and dissimilarity between trajectory pseudotimes}{63}{subsection.3.2.5}%
\contentsline {subsection}{\numberline {3.2.6}dynchro scales well to large datasets using representative pseudocells}{65}{subsection.3.2.6}%
\contentsline {section}{\numberline {3.3}Conclusion}{66}{section.3.3}%
\contentsline {section}{\numberline {3.4}Discussion}{68}{section.3.4}%
\contentsline {section}{\numberline {3.5}Methods}{69}{section.3.5}%
\contentsline {subsection}{\numberline {3.5.1}Lineages}{69}{subsection.3.5.1}%
\contentsline {subsubsection}{\nonumberline Construction}{69}{subsubsection*.62}%
\contentsline {subsubsection}{\nonumberline Comparison}{70}{subsubsection*.64}%
\contentsline {subsection}{\numberline {3.5.2}Dynamic time warping}{70}{subsection.3.5.2}%
\contentsline {subsection}{\numberline {3.5.3}Calculating average cost on path}{72}{subsection.3.5.3}%
\contentsline {subsection}{\numberline {3.5.4}Comparing multiple lineages}{72}{subsection.3.5.4}%
\contentsline {subsection}{\numberline {3.5.5}Interpolating cells}{72}{subsection.3.5.5}%
\contentsline {subsubsection}{\nonumberline Construction of pseudocells}{72}{subsubsection*.66}%
\contentsline {subsubsection}{\nonumberline Smoothing}{73}{subsubsection*.68}%
\contentsline {subsubsection}{\nonumberline Mapping back results}{73}{subsubsection*.70}%
\gdef \the@ipfilectr {}
\contentsline {part}{\numberline {II}Interoperability}{77}{part.2}%
\gdef \the@ipfilectr {@-9}
\contentsline {chapter}{\numberline {4}Polyglot programming for single-cell analysis}{81}{chapter.4}%
\contentsline {section}{\numberline {4.1}Introduction}{82}{section.4.1}%
\contentsline {subsection}{\numberline {4.1.1}Programming languages and software ecosystems}{84}{subsection.4.1.1}%
\contentsline {subsection}{\numberline {4.1.2}Interoperability strategies}{84}{subsection.4.1.2}%
\contentsline {paragraph}{\nonumberline Code porting}{85}{paragraph*.74}%
\contentsline {paragraph}{\nonumberline In-memory interoperability}{85}{paragraph*.76}%
\contentsline {paragraph}{\nonumberline Disk-based interoperability}{85}{paragraph*.78}%
\contentsline {paragraph}{\nonumberline Workflow frameworks}{85}{paragraph*.80}%
\contentsline {section}{\numberline {4.2}In-memory interoperability}{86}{section.4.2}%
\contentsline {paragraph}{\nonumberline Approach}{86}{paragraph*.82}%
\contentsline {subsection}{\numberline {4.2.1}Differences in programming language design}{87}{subsection.4.2.1}%
\contentsline {subsubsection}{\nonumberline Indexing: 0-based or 1-based}{87}{subsubsection*.84}%
\contentsline {subsubsection}{\nonumberline Column major vs row major matrices}{88}{subsubsection*.87}%
\contentsline {subsubsection}{\nonumberline Built-in data types: integers}{89}{subsubsection*.90}%
\contentsline {subsubsection}{\nonumberline Dots in variable names}{90}{subsubsection*.93}%
\contentsline {subsection}{\numberline {4.2.2}\software {rpy2}: a Python foreign function interface to R}{91}{subsection.4.2.2}%
\contentsline {subsubsection}{\nonumberline General usage}{91}{subsubsection*.95}%
\contentsline {subsubsection}{\nonumberline rpy2 extensions for single-cell analysis}{92}{subsubsection*.98}%
\contentsline {subsubsection}{\nonumberline Usecase: running DESeq2 in Python}{95}{subsubsection*.103}%
\contentsline {subsection}{\numberline {4.2.3}\software {reticulate}: an R foreign function interface to Python}{96}{subsection.4.2.3}%
\contentsline {subsubsection}{\nonumberline General usage}{96}{subsubsection*.105}%
\contentsline {subsubsection}{\nonumberline \software {reticulate} for single-cell analysis}{98}{subsubsection*.110}%
\contentsline {section}{\numberline {4.3}Disk-based interoperability}{101}{section.4.3}%
\contentsline {subsection}{\numberline {4.3.1}Interoperable file formats}{101}{subsection.4.3.1}%
\contentsline {subsubsection}{\nonumberline File format features}{102}{subsubsection*.113}%
\contentsline {paragraph}{\nonumberline Sparse matrix support}{102}{paragraph*.115}%
\contentsline {paragraph}{\nonumberline Large image support}{103}{paragraph*.117}%
\contentsline {paragraph}{\nonumberline Lazy chunk loading}{103}{paragraph*.119}%
\contentsline {paragraph}{\nonumberline Remote storage}{103}{paragraph*.121}%
\contentsline {subsection}{\numberline {4.3.2}Disk-based pipelines}{104}{subsection.4.3.2}%
\contentsline {subsubsection}{\nonumberline Pipelines with different environments}{105}{subsubsection*.125}%
\contentsline {subsubsection}{\nonumberline Containerized pipelines}{105}{subsubsection*.127}%
\contentsline {section}{\numberline {4.4}Production-ready workflows}{109}{section.4.4}%
\contentsline {subsection}{\numberline {4.4.1}Key qualities for workflows}{110}{subsection.4.4.1}%
\contentsline {subsection}{\numberline {4.4.2}Enabling technologies}{111}{subsection.4.4.2}%
\contentsline {subsection}{\numberline {4.4.3}Example workflow}{112}{subsection.4.4.3}%
\contentsline {section}{\numberline {4.5}Conclusion}{112}{section.4.5}%
\contentsline {paragraph}{\nonumberline One-off exploratory data analysis}{112}{paragraph*.131}%
\contentsline {paragraph}{\nonumberline Standard publication-focused analysis}{112}{paragraph*.133}%
\contentsline {paragraph}{\nonumberline Cohort-scale data with periodically added data}{113}{paragraph*.135}%
\contentsline {paragraph}{\nonumberline Benchmarking study}{113}{paragraph*.137}%
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-10}
\contentsline {chapter}{\numberline {5}anndataR improves interoperability between R and Python in single-cell transcriptomics}{117}{chapter.5}%
\contentsline {section}{\numberline {5.1}Introduction}{119}{section.5.1}%
\contentsline {subsection}{\numberline {5.1.1}Differences in structure}{119}{subsection.5.1.1}%
\contentsline {subsection}{\numberline {5.1.2}Different programming languages}{121}{subsection.5.1.2}%
\contentsline {section}{\numberline {5.2}Results}{124}{section.5.2}%
\contentsline {section}{\numberline {5.3}Software Design}{125}{section.5.3}%
\contentsline {section}{\numberline {5.4}Conclusion}{126}{section.5.4}%
\contentsline {section}{\numberline {5.5}Acknowledgments}{126}{section.5.5}%
\gdef \the@ipfilectr {}
\contentsline {part}{\numberline {III}Visualization}{135}{part.3}%
\gdef \the@ipfilectr {@-11}
\contentsline {chapter}{\numberline {6}funkyheatmap: Visualising data frames with mixed data types}{137}{chapter.6}%
\contentsline {section}{\numberline {6.1}Summary}{138}{section.6.1}%
\contentsline {section}{\numberline {6.2}Statement of need}{138}{section.6.2}%
\contentsline {section}{\numberline {6.3}Functionality}{139}{section.6.3}%
\contentsline {subsection}{\numberline {6.3.1}Example usage}{140}{subsection.6.3.1}%
\contentsline {section}{\numberline {6.4}Conclusion}{141}{section.6.4}%
\contentsline {section}{\numberline {6.5}Acknowledgements}{143}{section.6.5}%
\contentsline {section}{\numberline {6.6}Author Contributions}{143}{section.6.6}%
\contentsline {paragraph}{\nonumberline Common core}{144}{paragraph*.153}%
\contentsline {paragraph}{\nonumberline Code style}{144}{paragraph*.155}%
\contentsline {paragraph}{\nonumberline Testing}{145}{paragraph*.157}%
\contentsline {paragraph}{\nonumberline Diverging features}{145}{paragraph*.159}%
\gdef \the@ipfilectr {}
\contentsline {part}{\numberline {IV}Reflections}{149}{part.4}%
\gdef \the@ipfilectr {@-12}
\contentsline {chapter}{\numberline {7}Perspectives}{151}{chapter.7}%
\contentsline {section}{\numberline {7.1}Too many TI tools, too little guidance}{151}{section.7.1}%
\contentsline {paragraph}{\nonumberline dynchro aids in exploring trajectories}{153}{paragraph*.163}%
\contentsline {section}{\numberline {7.2}Interoperability is a community project}{153}{section.7.2}%
\contentsline {paragraph}{\nonumberline Community buy-in}{153}{paragraph*.165}%
\contentsline {paragraph}{\nonumberline Community needs}{154}{paragraph*.167}%
\contentsline {paragraph}{\nonumberline Community coordination}{155}{paragraph*.171}%
\contentsline {paragraph}{\nonumberline The future of anndataR}{156}{paragraph*.173}%
\contentsline {section}{\numberline {7.3}We must not overstate the importance of benchmarks}{156}{section.7.3}%
\contentsline {subsubsection}{\nonumberline Datasets}{156}{subsubsection*.175}%
\contentsline {subsubsection}{\nonumberline Ground truth}{157}{subsubsection*.177}%
\contentsline {subsubsection}{\nonumberline Metric selection}{157}{subsubsection*.179}%
\contentsline {subsubsection}{\nonumberline Other factors influence method selection}{158}{subsubsection*.181}%
\contentsline {subsubsection}{\nonumberline Changing best practices}{158}{subsubsection*.183}%
\gdef \the@ipfilectr {}
\gdef \the@ipfilectr {@-13}
\contentsline {chapter}{\numberline {8}Discussion}{163}{chapter.8}%
\contentsline {section}{\numberline {8.1}What is bioinformatics?}{163}{section.8.1}%
\contentsline {subsection}{\numberline {8.1.1}On bioinformatics}{163}{subsection.8.1.1}%
\contentsline {subsection}{\numberline {8.1.2}On data science}{165}{subsection.8.1.2}%
\contentsline {subsection}{\numberline {8.1.3}On science}{167}{subsection.8.1.3}%
\contentsline {subsection}{\numberline {8.1.4}To conclude}{169}{subsection.8.1.4}%
\contentsline {section}{\numberline {8.2}In practice}{169}{section.8.2}%
\contentsline {subsection}{\numberline {8.2.1}Code reproducibility}{170}{subsection.8.2.1}%
\contentsline {paragraph}{\nonumberline Replication difficulties}{171}{paragraph*.186}%
\contentsline {subsubsection}{\nonumberline Organise your code}{171}{subsubsection*.188}%
\contentsline {subsubsection}{\nonumberline Provide intermediary results}{171}{subsubsection*.190}%
\contentsline {subsubsection}{\nonumberline Keep track of dependency versions}{172}{subsubsection*.192}%
\contentsline {subsubsection}{\nonumberline Publish the code on a public repository}{172}{subsubsection*.194}%
\contentsline {subsection}{\numberline {8.2.2}Software maintainability}{172}{subsection.8.2.2}%
\contentsline {subsubsection}{\nonumberline Plan for maintenance}{173}{subsubsection*.196}%
\contentsline {subsubsection}{\nonumberline Carefully consider data formats}{173}{subsubsection*.198}%
\contentsline {subsubsection}{\nonumberline Limit dependencies}{174}{subsubsection*.200}%
\contentsline {subsection}{\numberline {8.2.3}Software maintenance is a shared responsibility}{174}{subsection.8.2.3}%
\contentsline {subsubsection}{\nonumberline Research labs and institutes}{174}{subsubsection*.202}%
\contentsline {subsubsection}{\nonumberline Consortia and communities}{175}{subsubsection*.204}%
\contentsline {subsubsection}{\nonumberline Funding agencies}{176}{subsubsection*.206}%
\contentsline {subsubsection}{\nonumberline Journals}{177}{subsubsection*.208}%
\contentsline {subsubsection}{\nonumberline Companies}{177}{subsubsection*.210}%
\contentsline {section}{\numberline {8.3}Standardization}{179}{section.8.3}%
\contentsline {subsection}{\numberline {8.3.1}What is a standard?}{179}{subsection.8.3.1}%
\contentsline {subsection}{\numberline {8.3.2}Open source organization}{180}{subsection.8.3.2}%
\contentsline {subsection}{\numberline {8.3.3}Open standards}{181}{subsection.8.3.3}%
\gdef \the@ipfilectr {}
\contentsline {part}{\numberline {V}Appendices}{191}{part.5}%
\gdef \the@ipfilectr {@-14}
\contentsline {chapter}{\numberline {A}Example workflow}{193}{appendix.A}%
\contentsline {section}{\numberline {A.1}Building the pipeline}{193}{section.A.1}%
\contentsline {subsection}{\numberline {A.1.1}Component: loading data (in Bash)}{193}{subsection.A.1.1}%
\contentsline {subsection}{\numberline {A.1.2}Component: select data (in Python)}{195}{subsection.A.1.2}%
\contentsline {subsection}{\numberline {A.1.3}Component: differential expression (in R)}{198}{subsection.A.1.3}%
\contentsline {subsection}{\numberline {A.1.4}Nextflow workflow}{202}{subsection.A.1.4}%
\contentsline {section}{\numberline {A.2}Running the workflow}{204}{section.A.2}%
\gdef \the@ipfilectr {}
\contentsline {chapter}{\numberline {B}Getting started with funkyheatmap}{207}{appendix.B}%
\providecommand \tocbasic@end@toc@file {}\tocbasic@end@toc@file