\@writefile{lof}{\contentsline{figure}{\numberline{7.1}{\ignorespaces\textbf{Illustration of the expectation-maximization algorithms}\textbf{A} illustration of ChIPPartitioning, an algorithm dedicated to the discovery of over-represented chromatin patterns, as described in \citep{nair_probabilistic_2014}. \textbf{B} illustration of EMSequence, an algorithm to discover over-represented DNA motifs. The overall design is the same. Both algorithms model the data has having being sampled from a distribution and perform a maximum-likelihood estimation of the distribution parameters from the data through an iterative procedure. EMJoint algorithm is the combination of both ChIPPartitioning and EMSequence at the same time.\relax}}{88}{figure.caption.35}}
\newlabel{atac_seq_em}{{7.1}{88}{\textbf{Illustration of the expectation-maximization algorithms}\textbf{A} illustration of ChIPPartitioning, an algorithm dedicated to the discovery of over-represented chromatin patterns, as described in \citep{nair_probabilistic_2014}. \textbf{B} illustration of EMSequence, an algorithm to discover over-represented DNA motifs. The overall design is the same. Both algorithms model the data has having being sampled from a distribution and perform a maximum-likelihood estimation of the distribution parameters from the data through an iterative procedure.\\ EMJoint algorithm is the combination of both ChIPPartitioning and EMSequence at the same time.\relax}{figure.caption.35}{}}
\citation{nair_probabilistic_2014}
\citation{nair_probabilistic_2014}
\citation{nair_probabilistic_2014}
\@writefile{toc}{\contentsline{subsubsection}{without shift and flip}{89}{figure.caption.35}}
\newlabel{atac_seq_emseq_likelihood}{{7.1}{89}{without shift and flip}{equation.7.5.1}{}}
\newlabel{atac_seq_emseq_update_model}{{7.2}{89}{without shift and flip}{equation.7.5.2}{}}
\@writefile{toc}{\contentsline{subsubsection}{with shift and flip}{89}{equation.7.5.2}}
\citation{nair_probabilistic_2014}
\citation{nair_probabilistic_2014}
\newlabel{atac_seq_emseq_likelihood_shift_flip}{{7.3}{90}{with shift and flip}{equation.7.5.3}{}}
\newlabel{atac_seq_emseq_reverse_motif}{{7.4}{90}{with shift and flip}{equation.7.5.4}{}}
\newlabel{atac_seq_emseq_update_model_shift_flip}{{7.5}{90}{with shift and flip}{equation.7.5.5}{}}
\@writefile{toc}{\contentsline{subsection}{\numberline{7.7.1}Aligning the binding sites}{93}{subsection.7.7.1}}
\@writefile{lof}{\contentsline{figure}{\numberline{7.2}{\ignorespaces\textbf{Central parts of the extended sequence and chromatin models} found in monocytes regulatory regions. Each read density and sequence pattern is a soft aggregation plot. The displayed logos correspond to the sequence class models found by EMSequence. The corresponding chromatin accessibility (red) and nucleosome occupancy (blue) are displayed atop of the logos. The classes are displayed by overall decreasing probability. A zoom over the central part of each class aggregation is shown in the top right inlet.\relax}}{94}{figure.caption.36}}
\newlabel{atac_seq_23class}{{7.2}{94}{\textbf{Central parts of the extended sequence and chromatin models} found in monocytes regulatory regions. Each read density and sequence pattern is a soft aggregation plot. The displayed logos correspond to the sequence class models found by EMSequence. The corresponding chromatin accessibility (red) and nucleosome occupancy (blue) are displayed atop of the logos. The classes are displayed by overall decreasing probability. A zoom over the central part of each class aggregation is shown in the top right inlet.\relax}{figure.caption.36}{}}
\@writefile{lof}{\contentsline{figure}{\numberline{7.3}{\ignorespaces\textbf{Soft aggregation plots of CTCF sub-classes} obtained by extracting CTCF class data and subjecting them to a ChIPPartitioning classification into 8 classes. The displayed logos correspond to each class sequence aggregation. The corresponding chromatin accessibility (red) and nucleosome occupancy (blue) are displayed atop of the logos. The classes are displayed by overall decreasing probability. A zoom over the central part of each class aggregation is shown in the top right inlet.\relax}}{96}{figure.caption.37}}
\newlabel{atac_seq_ctcf_subclass}{{7.3}{96}{\textbf{Soft aggregation plots of CTCF sub-classes} obtained by extracting CTCF class data and subjecting them to a ChIPPartitioning classification into 8 classes. The displayed logos correspond to each class sequence aggregation. The corresponding chromatin accessibility (red) and nucleosome occupancy (blue) are displayed atop of the logos. The classes are displayed by overall decreasing probability. A zoom over the central part of each class aggregation is shown in the top right inlet.\relax}{figure.caption.37}{}}
\@writefile{toc}{\contentsline{subsection}{\numberline{7.10.5}Extracting data assigned to a class}{100}{subsection.7.10.5}}
\newlabel{encode_peaks_algo_ndr_extend}{{2}{102}{Extracting data assigned to a class}{algocfline.2}{}}
\@writefile{loa}{\contentsline{algocf}{\numberline{2}{\ignorespaces Computes a matrix containing the data assigned to a given class $S$.\relax}}{102}{algocf.2}}
\newlabel{atac_seq_algo_extract_class}{{2}{102}{Extracting data assigned to a class}{algocf.2}{}}
\@writefile{toc}{\contentsline{subsection}{\numberline{7.10.9}Binding site prediction}{105}{subsection.7.10.9}}
\newlabel{atac_seq_method_pwmscan}{{7.10.9}{105}{Binding site prediction}{subsection.7.10.9}{}}
\citation{castro-mondragon_rsat_2017}
\@writefile{toc}{\contentsline{subsection}{\numberline{7.10.10}Realignment using JASPAR motifs}{106}{subsection.7.10.10}}
\@writefile{lot}{\contentsline{table}{\numberline{7.1}{\ignorespaces\textbf{TF binding models} from JASPAR matrix clustering. Each model can be retrieved within JASPAR matrix clustering (\url{http://jaspar2018.genereg.net/matrix-clusters/vertebrates/?detail=true}) using the cluster and node ID. "TFs covered" refers to all TF which models are children of the given node. "Name" refers to the label this model is referred to in the text and figures.\relax}}{107}{table.caption.38}}
\newlabel{atac_seq_motif_table}{{7.1}{107}{\textbf{TF binding models} from JASPAR matrix clustering. Each model can be retrieved within JASPAR matrix clustering (\url{http://jaspar2018.genereg.net/matrix-clusters/vertebrates/?detail=true}) using the cluster and node ID. "TFs covered" refers to all TF which models are children of the given node. "Name" refers to the label this model is referred to in the text and figures.\relax}{table.caption.38}{}}