ruby-minigraph 0.0.20.0

data/ext/minigraph/tex/minigraph.tex

@@ -0,0 +1,986 @@
+%% BioMed_Central_Tex_Template_v1.06
+
+\documentclass[twocolumn]{bmcart}
+
+%%% Load packages
+\usepackage{amsthm,amsmath}
+\RequirePackage{hyperref}
+\usepackage[utf8]{inputenc} %unicode support
+
+\usepackage{graphicx}
+%\def\includegraphic{}
+%\def\includegraphics{}
+
+%%% Put your definitions there:
+\startlocaldefs
+\endlocaldefs
+
+
+%%% Begin ...
+\begin{document}
+
+%%% Start of article front matter
+\begin{frontmatter}
+
+\begin{fmbox}
+\dochead{Method}
+
+\title{The design and construction of reference pangenome graphs with minigraph}
+
+\author[
+   addressref={aff1,aff2},          % id's of addresses, e.g. {aff1,aff2}
+   corref={aff1},                   % id of corresponding address, if any
+   email={hli@ds.dfci.harvard.edu}  % email address
+]{\inits{HL}\fnm{Heng} \snm{Li}}
+\author[
+   addressref={aff1,aff2},
+]{\inits{XF}\fnm{Xiaowen} \snm{Feng}}
+\author[
+   addressref={aff2},
+]{\inits{CC}\fnm{Chong} \snm{Chu}}
+
+\address[id=aff1]{%                           % unique id
+  \orgname{Department of Data Sciences, Dana-Farber Cancer Institute}, % university, etc
+  \city{Boston, MA 02215},                    % city
+  \cny{USA}                                   % country
+}
+\address[id=aff2]{%
+  \orgname{Department of Biomedical Informatics, Harvard Medical School},
+  \city{Boston, MA 02215},
+  \cny{USA}
+}
+
+\begin{abstractbox}
+
+\begin{abstract} % abstract
+The recent advances in sequencing technologies enable the assembly of
+individual genomes to the quality of the reference genome. How to integrate
+multiple genomes from the same species and make the integrated representation
+accessible to biologists remains an open challenge. Here, we propose a
+graph-based data model and associated formats to represent multiple genomes
+while preserving the coordinate of the linear reference genome. We implement
+our ideas in the minigraph toolkit and demonstrate that we can efficiently
+construct a pangenome graph and compactly encode tens of thousands of
+structural variants missing from the current reference genome.
+\end{abstract}
+
+\begin{keyword}
+\kwd{bioinformatics}
+\kwd{genomics}
+\kwd{pangenome}
+\end{keyword}
+
+\end{abstractbox}
+
+\end{fmbox}
+
+\end{frontmatter}
+
+%%
+\section*{Background}
+
+The human reference genome is a fundamental resource for human genetics and
+biomedical research. The primary sequences of the reference genome
+GRCh38~\cite{Schneider:2017aa} are a mosaic of haplotypes with each haplotype segment derived
+from a single human individual. They cannot represent the genetic diversity in
+human populations and as a result, each individual may carry thousands of large
+germline variants absent from the reference genome~\cite{Huddleston:2017aa}.
+Some of these variants are likely associated with phenotype~\cite{Eichler_2010}
+but are often missed or misinterpreted when we map sequence data to GRCh38, in
+particular with short reads~\cite{Li:2018aa}. This under-representation of
+genetic diversity may become a limiting factor in our understanding of genetic
+variations.
+
+Meanwhile, the advances in long-read sequencing technologies make it possible
+to assemble a human individual to a quality comparable to
+GRCh38~\cite{Schneider:2017aa,Wenger_2019}. There are already a dozen of
+high-quality human assemblies available in GenBank~\cite{Audano:2019aa}.
+Properly integrating these genomes into a reference \emph{pangenome}, which
+refers to a collection of genomes~\cite{cpgc:2016aa}, would potentially address
+the issues with a single linear reference.
+
+A straightforward way to represent a pangenome is to store unaligned genomes
+in a full-text index that compresses redundancies in sequences identical
+between individuals~\cite{Makinen:2010aa,Liu_2016,Boucher_2019}. We may
+retrieve individual genomes from the index, inspect the k-mer spectrum and test
+the presence of k-mers using standard techniques. In principle, it is also
+possible to apply canonical read alignment algorithms to map sequences to
+the collection, but in practice, the redundant hits to multiple genomes will
+confuse downstream mapping-based analyses~\cite{NA2016159}. It is not clear how
+to resolve these multiple mappings.
+
+The other class of methods encodes multiple genomes into a sequence graph,
+usually by collapsing identical or similar sequences between genomes onto a
+single representative sequence. The results in a \emph{pangenome graph}. A
+pangenome graph is a powerful tool to identify core genome, the part of a
+genome or gene set that is shared across the majority of the strains or related species
+in a clade~\cite{Vernikos:2015aa}. A common way to construct a basic pangenome
+graph is to generate a compacted de Bruijn graph
+(cDBG)~\cite{Marcus:2014xy,Baier_2015,Beller:2016ab,Chikhi:2015aa,Minkin_2016,Chikhi_2016,almodaresi_et_al:LIPIcs:2017:7657}
+from a set of genomes. Basic cDBG does not keep sample information.
+\cite{Iqbal:2012aa} proposed colored cDBG with each color represents a sample
+or a population. Colored cDBG can be constructed
+efficiently~\cite{Muggli_2019,Holley695338}. However, a colored cDBG discards
+the chromosomal coordinate and thus disallows the mapping of genomic features.
+It often includes connections absent from the input genomes and thus encodes
+sequences more than the input. A colored cDBG cannot serve as a
+\emph{reference} pangenome graph, either.  deBGA~\cite{Liu:2016ac} addresses
+the issue by labeling each unitig with its possibly multiple locations in the
+input genome(s). Pufferfish~\cite{Almodaresi:2018aa} further reduces its space
+requirement. Nonetheless, given hundreds of human genomes, there will be many
+more vertices in the graph and most vertices are associated with hundreds of
+labels. Whether deBGA and pufferfish can scale to such datasets remains an open
+question. GBWT~\cite{Sir_n_2019} provides another practical solution to storage
+and indexing, but no existing tools can practically construct a cDBG for many
+human genomes in the GBWT representation.
+
+In addition to cDBG, we can derive a reference pangenome
+graph from a single linear multi-sequence alignment (MSA)~\cite{Dilthey_2015,Dilthey_2019}.
+It has been used for HLA typing but is not applicable to whole chromosomes when
+they cannot be included in a single linear MSA. The third and possibly the most
+popular approach to reference graph generation is to call variants from other
+sources and then incorporate these variants, often in the VCF format~\cite{Danecek:2011qy}, into
+the reference genome as alternative
+paths~\cite{Eggertsson:2017aa,Rakocevic_2019,Sibbesen:2018aa,Biederstedt:2018aa,Eggertsson_2019}.
+However, because VCF does not define coordinates on insertions, this approach
+cannot properly encode variations on long insertions and is therefore limited
+to simple variations. There are no satisfactory solutions to the construction
+of reference pangenome graphs.
+
+In this article, we introduce the reference Graphical Fragment Assembly (rGFA)
+format to model reference pangenome graphs. We propose and demonstrate an
+incremental procedure to construct graphs under this model. The resulting
+graphs encode structural variations (SVs) of length 100bp or longer without haplotype
+information.  Our implementation, minigraph~\cite{Li_minigraph:2020aa}
+(\href{https://github.com/lh3/minigraph}{https://github.com/lh3/minigraph}),
+can construct a pangenome graph from twenty human assemblies in three hours.
+
+\section*{Results}
+
+We will first describe a data model for reference pangenome graphs, which
+establishes the foundation of this article. We will then present a new
+sequence-to-graph mapper, minigraph, and show how this mapper incrementally
+constructs a pangenome graph. We will demonstrate the utility of pangenome
+graphs with a human graph generated from twenty human haplotypes and a primate
+graph generated from four species.
+
+\subsection*{Modeling reference pangenome graphs}
+
+\subsubsection*{Sequence graphs}
+
+There are several equivalent ways to define a sequence graph. In this article,
+a \emph{sequence graph} $G(V,E)$ is a bidirected graph. Each vertex $v\in V$ is
+associated with a DNA sequence; each edge $e\in E$ has two directions, one for
+each endpoint, which leads to four types of edges: forward-forward,
+reverse-forward, forward-reverse and reverse-reverse. The directions on an edge
+dictate how a sequence is spelled from a walk/path in the graph. Common
+assembly graphs, such as the overlap graph, string graph and de Bruijn graph
+can all be formulated as sequence graphs.
+
+\begin{figure}[t]
+\includegraphics[width=.47\textwidth]{Fig1}
+\caption{\csentence{Example rGFA and GAF formats.} {\bf (a)} Example rGFA
+  format. rGFA-specific tags include SN, name of the stable sequence from which
+  the vertex is derived; SO, offset on the stable sequence; SR, rank: 0 if the
+  vertex or edge is on the linear reference; $>$0 for non-reference.  {\bf (b)}
+  Corresponding sequence graph. Each thick arrow represents an oriented DNA
+  sequence. {\bf (c)} Example GAF format, using the segment coordinate, for
+  reads ``${\tt GTGGCT}$'' and ``${\tt CGTTTCC}$'' mapped to the graph. {\bf
+  (d)} Equivalent GAF format using the stable coordinate.}\label{fig:rgfa}
+\end{figure}
+
+The Graphical Fragment Assembly (GFA) format~\cite{Li:2016aa} describes
+sequence graphs. The core of GFA is defined by the following grammar:
+
+{\footnotesize
+\begin{verbatim}
+
+  <GFA>     <- (<segment> | <link>)+
+  <segment> <- `S' <segId> <segSeq>
+  <link>    <- `L' <segId> [+-] <segId> [+-] <cigar>
+
+\end{verbatim}}
+
+{\flushleft
+A line starting with letter ``${\tt S}$'' corresponds to a vertex and a line
+starting with ``${\tt L}$'' corresponds
+to a bidirected edge. In a de Bruijn graph, we often attach sequences to edges
+instead of vertices~\cite{Pevzner:2001vn,Gnerre:2011ys}. To avoid the confusion, in this
+article, we also call a vertex as a \emph{segment} and call an edge as a
+\emph{link}, following the GFA terminology.  Fig.~\ref{fig:rgfa}a shows an
+example GFA that encodes Fig.~\ref{fig:rgfa}b.
+}
+
+A sequence graph in the GFA format natively defines a \emph{segment coordinate}
+system where each base in the graph is uniquely indexed by a
+2-tuple $({\rm segId},{\rm segOffset})$. For example, in
+Fig~\ref{fig:rgfa}a, the base at position $({\rm s2},2)$ is ``{\tt G}''.
+A major problem with this coordinate is that it is decoupled from linear
+annotations and is sensitive to graph transformations. For example, if we split
+a segment into two connected segments, the set of sequences spelled from the graph
+remains the same, but the segment coordinates will be changed. Due to the
+instability of segment coordinate, a basic sequence graph is inadequate for a
+reference graph.
+
+\subsubsection*{Reference pangenome graphs}
+
+We propose the reference GFA (rGFA) format to encode reference pangenome graphs.
+rGFA is an extension to GFA with three additional tags that indicate the origin
+of a segment from linear genomes (Fig.~\ref{fig:rgfa}a). This simple addition
+gives us a unique stable coordinate system as an extension to the linear
+reference coordinate (e.g. GRCh38). We can pinpoint a position such as
+``{\sf chr1:9}'' in the graph and map existing annotations onto the graph. We can
+also report a path or walk in the stable coordinate. For example, path
+``{\sf s1$\to$s2$\to$s3}'' unambiguously corresponds to ``{\sf
+chr1:0-5$\to$chr1:5-8$\to$chr1:8-12}'' or simply ``{\sf chr1:0-12}'' if we
+merge adjacent coordinate; similarly, ``{\sf s1$\to$s2$\to$s5$\to$s6}''
+corresponds to ``{\sf chr1:0-8$\to$foo:8-16}''. We will formally describe the
+path format when introducing the GAF format in the next section.
+
+In rGFA, each segment is associated with one origin. This apparently trivial
+requirement in fact imposes a strong restriction on the types of graphs rGFA
+can encode: it forbids the collapse of different regions from one sequence,
+which would often happen in a cDBG. We consider this restriction an
+advantage of rGFA because it requires the graph to have a ``linear'' flavor
+intuitively and simplifies the data structure to store the graph.
+
+For simplicity, rGFA disallows overlaps between edges and forbids multiple
+edges (more than one edges between the same pair of vertices). These two
+restrictions help to avoid ambiguity and reduce the complexity in
+implementation. They are not strictly necessary in theory.
+
+\subsubsection*{The Graphical mApping Format (GAF)}
+
+\begin{table}[tb]
+\caption{The Graphical mApping Format (GAF)}\label{tab:gaf}
+\begin{tabular}{rcp{6cm}}
+\hline
+Col & Type  & Description \\ \hline
+1  & string & Query sequence name \\
+2  & int    & Query sequence length \\
+3  & int    & Query start coordinate (0-based; closed) \\
+4  & int    & Query end coordinate (0-based; open) \\
+5  & char   & Strand relative to col. 6 \\
+6  & string & Graph path matching regular expression \texttt{/([><][\char94\char92s><]+(:\char92d+-\char92d+)?)+\char124([\char94\char92s><]+)/}\\
+7  & int    & Path sequence length \\
+8  & int    & Path start coordinate \\
+9  & int    & Path end coordinate \\
+10 & int    & Number of matching bases in the mapping \\
+11 & int    & Number of bases, including gaps, in the mapping \\
+12 & int    & Mapping quality (0--255 with 255 for missing) \\ \hline
+\end{tabular}
+\end{table}
+
+As there are no text formats for sequence-to-graph alignment, we propose a new
+Graphical mApping Format (GAF) by extending the Pairwise mApping Format
+(PAF)~\cite{Li:2016aa}. GAF is TAB-delimited with each column defined in
+Table~\ref{tab:gaf}. Column 6 encodes a path on the graph. It follows the
+formal grammar below:
+
+{\footnotesize
+\begin{verbatim}
+
+  <path>       <- <stableId> | <orientIntv>+
+  <orientIntv> <- (`>' | `<') (<segId> | <stableIntv>)
+  <stableIntv> <- <stableId> `:' <start> `-' <end>
+
+\end{verbatim}}
+
+{\flushleft
+In this grammar, {\tt <segId>} is a segment identifier on an S-line in rGFA;
+{\tt <stableId>} is a stable sequence name at the {\tt SN} tag on the
+corresponding S-line. Column 6 can be either a path in the segment coordinate
+(Fig.~\ref{fig:rgfa}c) or an equivalent path in the stable coordinate
+(Fig.~\ref{fig:rgfa}d). We can merge adjacent stable coordinates if the two
+segments are originated from the same stable sequence and the end offset of the
+first segment is equal to the start offset of the second segment. For example,
+``{\tt >chr1:0-5>chr1:5-8}'' can be simplified to ``{\tt >chr1:0-8}''.
+Furthermore, if a path in column 6 is derived from one reference sequence, we
+recommend to replace it with the entire reference path on the forward
+orientation (e.g. see ``read1'' in Fig.~\ref{fig:rgfa}d). With this convention,
+a GAF line is reduced to PAF for a sequence mapped to a reference sequence.
+Similar to PAF, GAF also allows optional tags in the SAM-like format. Base
+alignment is kept at the {\tt cg} tag.}
+
+Minigraph produces GAF in both the segment and the stable coordinate.
+GraphAligner~\cite{Rautiainen810812} produces GAF in the segment coordinate
+only, which can be converted to the stable coordinate.
+
+\begin{figure}[t]
+\includegraphics[width=.47\textwidth]{Fig2}
+\caption{\csentence{Minigraph algorithms.} {\bf (a)} Diagram of the minigraph
+  mapping algorithm. Minigraph seeds alignments with minimizers, finds good
+  enough linear chains, connects them in the graph and seeks the most weighted
+  path as a graph chain. {\bf (b)} Diagram of incremental graph construction. A
+  graph is iteratively constructed by mapping each assembly to an existing
+  graph and augmenting the graph with long poorly mapped sequences in the
+  assembly.}\label{fig:mg}
+\end{figure}
+
+\subsection*{Sequence-to-graph mapping}
+
+Our incremental graph construction algorithm relies on genome-to-graph
+alignment (Fig.~\ref{fig:mg}b). As existing sequence-to-graph
+aligners~\cite{Rautiainen810812,Garrison:2018aa} do not work with
+chromosome-long query sequences, we adapted minimap2~\cite{Li:2018ab} for our
+purpose and implemented minigraph (Fig.~\ref{fig:mg}a). Briefly, minigraph uses
+a minimap2-like algorithm to find local hits to segments in the graph, ignoring
+the graph topology. It then chains these local hits if they are connected on
+the graph, possibly through cycles. This gives the approximate mapping locations. Minigraph does not
+perform base-level alignment. This is because the graph we construct encodes
+SVs and rarely contains paths similar at the base level. The best mapping is
+often clear without base alignment.
+
+\begin{table}[b]
+\caption{Performance of sequence-to-graph mapping}\label{tab:mgvga}
+\begin{tabular}{lrr}
+\hline
+& minigraph & GraphAligner \\
+\hline
+Indexing time (wall-clock sec) & 100 & 589 \\
+Mapping time (wall-clock sec) & 79 & 140 \\
+Peak RAM (GB)          & 19.5 & 27.2 \\
+Percent unmapped reads & 0.5\% & 0\% \\
+Percent wrong mappings & 1.7\% & 4.6\% \\
+\hline
+\end{tabular}
+\end{table}
+
+To evaluate the accuracy of minigraph mapping, we simulated PacBio reads from
+GRCh38 with PBSIM~\cite{Ono:2013aa} and mapped them to the graph we constructed
+in the next section. Table~\ref{tab:mgvga} compares the performance of
+minigraph and GraphAligner~\cite{Rautiainen810812} v1.0.10 on 68,857 simulated
+reads mapped over 8 CPU threads. {\color{black} The N50 read length is 15kb.
+9,862 reads are mapped across two or more segments by GraphAligner. Note that
+both minigraph and GraphAligner ignore the stable coordinates during mapping.
+All segments, originated either from GRCh38 or from individual genomes, are
+treated equally. To this end, while we simulated reads from GRCh38, we are also
+evaluating how well mappers work with complex SVs present in any input
+samples.}
+
+On this dataset, minigraph
+is faster than GraphAligner and uses less memory, partly because minigraph does
+not perform base alignment.
+As is shown in Table~\ref{tab:mgvga}, minigraph is more accurate than
+GraphAligner. This is counter-intuitive given that GraphAligner does base
+alignment. Close inspection reveals that most mismapped reads by minigraph are
+mapped to the correct genomic loci but wrong graph paths. On the contrary, most
+mismapped reads by GraphAligner are mapped to wrong genomic loci. This suggests
+minigraph is better at finding approximate mapping locations but GraphAligner
+is better at disambiguating similar graph paths.  Combining the strength of
+both could lead to a better graph mapper. We do plan to implement base-level
+alignment in minigraph in future.
+
+We have also tried vg v1.21.0~\cite{Garrison:2018aa}. It indexed the same graph in 14.7 wall-clock
+hours and mapped the simulated reads in 1.8 hours over 8 threads, tens of times
+slower than minigraph and GraphAligner. However, no reads are mapped in the
+output. We have not been able to make vg work with our data.
+
+\subsection*{Generating pangenome graphs}
+
+Fig.~\ref{fig:mg}b shows how minigraph constructs a pangenome graph (see
+Methods for details). This procedure is similar to multiple sequence alignment
+via partial order graph~\cite{Lee_2002} except that minigraph works with cyclic
+graphs and ignores small variants. Minigraph only considers SVs of
+100bp--100kb in length and ignores SVs in alignments shorter than 100kb.
+For each input assembly, it filters out regions covered by two or more primary
+alignments longer than 20kb in the assembly. This filter avoids paralogous
+regions in a sample and guarantees that graphs generated by minigraph can be
+modeled by rGFA.
+
+As a sanity check, we compared minigraph to dipcall
+(\href{https://github.com/lh3/dipcall}{https://github.com/lh3/dipcall}) on
+calling SVs 100bp or longer from a synthetic diploid sample composed of CHM1
+and CHM13~\cite{Li:2018aa}. Given two SV callsets $A$ and $B$, we say a call in
+$A$ is \emph{missed} in callset $B$ if there are no calls in $B$ within 1000bp
+from the call in $A$. With this criterion, 2.7\% of 14,792 SVs called by
+dipcall are missed by minigraph; 6.0\% of 14,932 minigraph SVs are missed by
+dipcall. We manually inspected tens of differences in
+IGV~\cite{Robinson:2011aa} and identified two causes. First, an INDEL longer
+than 100bp called by one caller may be split into two shorter INDELs by the
+other caller. There are often more than one smaller SVs around a missed SV
+call. Second, dipcall skips regions involving high density of SNPs or involving
+both long insertions and long deletions, but minigraph connects these events
+and calls SVs in such regions. It tends to call more SVs. Overall, we believe
+minigraph and dipcall found similar sets of SVs.
+
+\begin{table}[tb]
+\caption{Assemblies used for graph construction}\label{tab:asm}
+\begin{tabular}{llll}
+\hline
+Name & Species & Population  & Accession/Source \\ \hline
+CHM1 & Human   & N/A         & GCA\_001297185.1 \\
+CHM13 & Human  & N/A         & GCA\_000983455.1 \\
+NA12878 & Human & European   & \cite{Garg810341}, phased \\
+NA24385 & Human & Jewish     & \cite{Garg810341}, phased \\
+PGP1 & Human & N/A           & \cite{Garg810341}, phased \\
+NA19240 & Human & African    & GCA\_001524155.4 \\
+HG00514 & Human & East Asian & GCA\_002180035.3 \\
+HG01352 & Human & American   & GCA\_002209525.2 \\
+NA19434 & Human & African    & GCA\_002872155.1 \\
+HG02818 & Human & African    & GCA\_003574075.1 \\
+HG03486 & Human & African    & GCA\_003086635.1 \\
+HG03807 & Human & South Asian& GCA\_003601015.1 \\
+HG00733 & Human & American   & GCA\_002208065.1 \\
+HG02059 & Human & East Asian & GCA\_003070785.1 \\
+HG00268 & Human & European   & GCA\_008065235.1 \\
+HG04217 & Human & South Asian& GCA\_007821485.1 \\
+AK1     & Human & East Asian & GCA\_001750385.1 \\
+Clint   & Chimpanzee &       & GCA\_002880755.3 \\
+Susie   & Gorilla &          & GCA\_900006655.3 \\
+Kamilah & Gorilla &          & GCA\_008122165.1 \\
+Susie   & Orangutan &        & GCA\_002880775.3 \\
+\hline
+\end{tabular}
+\end{table}
+
+\begin{figure*}[htbp]
+\includegraphics[width=.95\textwidth]{Fig3}
+\caption{\csentence{Characteristics of the human and the great ape graphs.} {\bf
+  (a)} Human variations stratified by repeat class and by the number of
+  alleles of each variation. The repeat annotation was obtained from the
+  longest allele of each variation. VNTR: variable-number tandem repeat, a
+  tandem repeat with the unit motif length $\ge$7bp. STR: short random repeat,
+  a tandem repeat with the unit motif length $\le$6bp. LCR: low-complexity
+  regions. Mixed-inter.: a variation involving $\ge$2 types of interspersed
+  repeats. {\bf (b)} Great ape variations stratified by repeat class and by the
+  number of alleles. {\bf (c)} Human biallelic variations stratified by repeat
+  class and by insertion to/deletion from GRCh38. Both alleles are required to
+  be covered in all assemblies. {\bf (d)} Human-specific biallelic variations
+  stratified by repeat class and by insertion to/deletion from GRCh38. Red bars
+  correspond to insertions to the human lineage. {\bf (e)} Distribution of
+  different types of human variations along chromosomes.  {\bf (f)} Boxplot of
+  the longest allele length in each repeat class. Outliers are omitted for the
+  clarity of the figure.}\label{fig:anno}
+\end{figure*}
+
+\subsection*{A human pangenome graph}
+
+Starting with GRCh38, we constructed a human pangenome graph from 20 human
+haplotypes or haplotype-collapsed assemblies (Table~\ref{tab:asm}). It took
+minigraph 2.7 wall-clock hours over 24 CPU threads to generate this graph. The
+peak memory is 98.1GB. The resulting graph consists of 148,618 segments and
+214,995 links. It contains 37,332 variations, where a \emph{variation}
+denotes a minimal subgraph that has a single source and a single sink with both
+segments coming from GRCh38. A path through the bubble between the source and
+and the sink represents an \emph{allele}.
+
+Variations in the human graph are enriched with Alus and VNTRs
+(Fig.~\ref{fig:anno}a). While interspersed repeats are about evenly distributed
+along chromosomes except in the pseudoautosomal regions (Fig.~\ref{fig:anno}e),
+VNTRs are enriched towards telomeres~\cite{Audano:2019aa}. It is worth noting
+the density of minisatellites is also higher in subtelomeres. If we normalize
+the density of VNTRs in the pangenome graph by the density of minisatellites in
+GRCh38, the enrichment of VNTRs towards telomeres is still visible but becomes
+less prominent. At the same time, repeat-less variations are also enriched
+towards the ends of chromosomes (green areas in Fig.~\ref{fig:anno}e),
+suggesting subtelomeres tend to harbor SVs anyway. We also
+identified 85 processed pseudogenes among these variations.
+
+\begin{figure}
+\includegraphics[width=.46\textwidth]{igv-edit.png}
+\caption{\csentence{IGV screenshot of a region enriched with long insertions.}
+  Numbers on wide purple bars indicate insertion lengths. CLR: PacBio noisy
+  continuous long reads. HiFi: PacBio high-fidelity reads.}\label{fig:igv}
+\end{figure}
+
+Another noticeable feature of VNTRs is that over half of VNTR variations are
+multiallelic (Fig.~\ref{fig:anno}a). Fig.~\ref{fig:igv} shows a multi-allelic
+region composed of VNTRs. We can see many insertions of different lengths. The
+two different NA12878 assemblies also disagree with each other, which we often
+see around other VNTR loci in NA12878 as well. We have not inspected raw reads
+in this particular example, but we tend to believe the disagreement is caused
+by local misassemblies rather than somatic mutations. In addition, due to the
+multiallelic nature of such VNTRs, the two haplotypes in a human individual are
+often different. Assemblies mixing the two haplotypes (aka collapsed
+assemblies) may have more troubles in these regions. Multiallelic VNTRs are
+hard to assemble correctly.
+
+Multiallelic VNTRs are also hard to align and to call. In Fig.~\ref{fig:igv},
+the insertion positions are often different, which could be caused by a few
+mutations or sequencing errors. A naive alignment-based SV caller would call a
+dozen of low-frequency insertions in this region, which does not reflect these
+correlated events. Without base-level alignment, minigraph may
+have more troubles with obtaining the optimal alignment in these complex VNTR
+regions. Improved data quality, assembly algorithms and graph mapping
+algorithms are required to investigate VNTR regions in detail.
+
+\subsection*{A great ape pangenome graph}
+
+We also constructed a great ape pangenome graph from GRCh38, one chimpanzee,
+two gorillas and one orangutan (Table~\ref{tab:asm}). This graph contains
+206,452 variations, over four times more than the human graph. About half of
+variations are originated from orangutan, the species most distant from human.
+
+In the great ape graph, the L1-to-Alu ratio is close to 1:1, much higher than
+the ratio in the human graph (Fig.~\ref{fig:anno}b vs Fig.~\ref{fig:anno}a).
+This is perhaps correlated with the elevated L1 activity in great
+apes~\cite{Mathews:2003aa}. Of retrotransposon-related variations specific to
+the human lineage, the overwhelming majority are insertions
+(Fig.~\ref{fig:anno}d), which is expected as transpositions lead to insertions
+only.  Most human-specific Alu deletions are incomplete and involve ancient Alu
+subfamilies. They are likely genomic deletions that happen to hit Alus. In
+contrast, the majority of ``partial-repeats'' are deletions from the human
+lineage. Two thirds of autosomal insertions in this category are segmental
+duplications in GRCh38. In all, minigraph is an efficient tool to study closely
+related species.
+
+\subsection*{Blacklist regions from human pangenome graphs}
+
+The human pangenome graph effectively encodes SVs $\ge$100bp
+in 20 genomes. These large-scale variations could be a frequent source of
+technical artifacts in variant calling with short reads. To test this
+hypothesis, we compared short-read SNP calls with vs without regions around SVs
+in the pangenome graph.
+
+We constructed a human pangenome graph excluding CHM1 and CHM13, the two
+samples used in the SynDip benchmark~\cite{Li:2018aa}, and generated regions
+around variations (see Methods), which we call as \emph{blacklist regions},
+following the rationale in~\cite{Amemiya:2019aa}.  Blacklist regions is totaled
+29.2Mb in length, intersecting 0.7\% of confident regions in
+SynDip~\cite{Li:2018aa}; 0.7\% of truth SNPs are contained in blacklist regions
+-- true SNPs are not enriched in blacklist regions.
+
+We mapped short reads used in~\cite{Li:2018aa} with minimap2 and called
+variants with GATK v4.1.2~\cite{Depristo:2011vn}. This callset
+contains 32,879 false positive SNPs, 21\% of which fall in blacklist regions --
+false SNP calls are highly enriched in this $<$1\% region of human genome. This
+confirms a noticeable fraction of false SNP calls using short reads are
+resulted from misalignment involving SVs.
+
+\section*{Discussion}
+
+Based on the GFA assembly format~\cite{Li:2016aa}, we proposed the rGFA format,
+which defines a data model for reference pangenome graphs at the same time.
+rGFA takes a linear reference genome as the backbone and maintains the
+conceptual ``linearity'' of input genomes.
+
+rGFA is not the only pangenome graph model. Vg~\cite{Garrison:2018aa}
+encodes a stable sequence with a path through the sequence graph~\cite{10.12688/f1000research.19630.1}. A segment
+in the graph may occur on multiple paths, or occur multiple times on one path
+if there are cycles in the graph. This way, vg allows different regions in one
+chromosome collapsed to one segment. We call such a graph as a collapsed graph. rGFA
+cannot encode a collapsed graph. The vg model is thus more general.
+
+In our view, however, the reference pangenome graph should not be a collapsed
+graph. In a collapsed graph, the definition of orthology is not clear because
+multiple sequences from the same sample may go through the same segment.
+Without the concept of orthology, we cannot define variations, either.  In
+addition, due to the one-to-many relationship between segments and the
+reference genome, it is intricate to derive the stable coordinate of a path in
+a collapsed graph. For example, suppose segment {\sf s1} corresponds to two
+regions {\sf chr1:100-200} and {\sf chr1:500-600}. To convert a path {\sf
+s2$\to$s1$\to$s3} to the stable coordinate, we have to inspect adjacent
+segments to tell which {\sf s1} corresponds to; this becomes more challenging
+when {\sf s2} and {\sf s3} represent multiple regions in the reference genome.
+In contrast, rGFA inherently forbids a collapsed graph and avoids the potential
+issues above. This makes rGFA simpler than vg's path model and easier to work
+with.
+
+To demonstrate practical applications of rGFA, we developed minigraph to
+incrementally generate pangenome graphs. It can generate a graph from 20
+genomes in three hours and can scale to hundreds of genomes in future. A
+limitation of minigraph is that it does not perform base alignment and may be
+confused by similar paths in the graph. {\color{black} Unfortunately, base-level
+sequence-to-graph alignment is not a fully solved problem. Partial-order graph
+alignment~\cite{Lee_2002} and PaSGAL~\cite{DBLP:conf/ipps/JainMZDA19} only work
+with directed acyclic graphs (DAGs). Vg~\cite{Garrison:2018aa} uses a heuristic
+to unroll cycles but it is expotential in time in the worst case and for DAGs,
+its exact mode is tens of times slower than PaSGAL. Antipov et
+al~\cite{Antipov:2016aa} proved that alignment against cyclic graphs can be
+done in polynomial time. GraphAligner~\cite{Rautiainen810812} implements a
+fast quadratic algorithm for computing edit distance~\cite{Rautiainen_2019}.
+However, edit distance based alignment disallows long INDELs and is often
+inadequate for accurate variant calling. Jain et al~\cite{Jain_2020} recently
+proposed a quadratic algorithm for alignment with affine gap penalty but the
+authors focused on the theoretical analysis only. To the best of our knowledge,
+no tools can efficiently perform sequence-to-graph alignment under affine gap
+cost. We plan to learn from the existing algorithms and implement fast base
+alignment in minigraph in future. This may take significant effort.}
+
+Another limitation of minigraph is
+that it is unable to align sequences against a graph encoding all small variants.
+Such a graph will be composed of millions of short segments. Not
+indexing minimizers across segments, minigraph will fail to seed the initial
+linear chains. This limitation can only be resolved by completely changing the
+minigraph mapping algorithm. Nonetheless, small variants are easier to
+analyze with the standard methods. Incorporating these variants unnecessarily
+enlarges the graph, complicates implementations, increases the rate of false
+mappings~\cite{Pritt_2018} and reduces the performance of common tasks. There
+is also no known algorithm that can construct such a complex graph for hundreds
+of human genomes.
+
+Minigraph does not keep track of the sample information as of now. To address
+this issue, we are considering to implement colored rGFA, similar to colored de
+Bruijn graphs~\cite{Iqbal:2012aa}. In a colored rGFA, a color represents one
+sample.  Each segment or link is associated with one or multiple colors,
+indicating the sources of the segment or the link. Colors can be stored in an
+rGFA tag or in a separate segment/link-by-sample binary
+matrix~\cite{Holley695338}. The matrix representation may be more compact given
+a large number of samples.
+
+We have shown minigraph can be a fast and powerful research tool to summarize
+SVs at the population scale and to study the evolution of closely related
+species. A more practical question is how a reference pangenome graph may
+influence routine data analysis. Here is our limited view.
+
+We think a critical role a reference graph plays is that it extends the
+coordinate system of a linear reference genome. This allows us to annotate
+variations in highly diverse regions such as the human HLA and KIR regions. The
+existing pipelines largely ignore these variations because most of them cannot
+be encoded in the primary assembly of GRCh38.
+
+The extended graph coordinate system further helps to consistently represent
+complex SVs. Given multiple samples, the current practice is to call SVs from
+individual samples and then merge them. Two subtly different SVs, especially
+long insertions, may be called at two distinct locations and treated as
+separate events. With the minigraph procedure, the two SVs are likely to
+be aligned together as long as they are similar to each other and are
+sufficiently different from the reference allele. To some extent, minigraph is
+performing multiple sequence alignment with partial order
+alignment~\cite{Lee_2002}. This procedure is more robust to different
+representations of the same SV than naive merging. When we refer to a SNP, we often use its
+chromosomal coordinate such as ``chr1:12345''. We rarely do so for SVs because
+their positions are sensitive to alignment and SV callers. The more consistent
+SV representation implied by a pangenome graph will help to alleviate the issue
+and subsequently facilitate the genotyping of
+SVs~\cite{Hickey_2020,Eggertsson_2019,Chen_2019}.
+
+While we believe a reference pangenome graph will make complex variations more
+accessible by geneticists and biologists, we suspect a great majority of
+biomedical researchers will still rely on a linear reference genome due to the
+conceptual simplicity of linear genomes and the mature tool chains developed in
+decades. Many analyses such as SNP calling in well behaved regions do not
+benefit much from a pangenome representation, either. Nonetheless, a pangenome
+reference still helps applications based on linear references. With a graph
+reference, we may blacklist regions enriched with SVs that lead to small variant
+calling errors.  We may potentially generate ``decoy'' sequences that are
+missing from the primary assembly to attract falsely mapped reads away. We may
+perform read alignment against a graph, project the alignment to the linear
+coordinate and finish the rest of analyses in the linear space. We anticipate a
+pangenome reference to supplement the linear reference, not to replace it.
+
+\section*{Conclusions}
+
+Complex human sequence variations are like genomic dark matter: they are
+pervasive in our genomes but are often opaque to the assay with the existing
+tools. We envision a pangenome graph reference will become an effective
+means to the study of these complex variations. We proposed a data model (rGFA),
+designed formats (rGFA and GAF) and developed companion tools (minigraph and
+gfatools) to demonstrate the feasibility of our vision. Our work is still
+preliminary but it is likely to set a starting point to the development of the
+next-generation graph-based tools, which may ultimately help us to understand
+our genomes better.
+
+\section*{Methods}
+
+\subsection*{The minigraph mapping algorithm}
+
+\subsubsection*{Seeding and linear chaining}
+Similar to minimap2, minigraph uses minimizers on segments as seeds. It also
+applies a similar chaining algorithm but with different scoring and with a new
+heuristic to speed up chaining over long distances. For the completeness of
+this article, we will describe part of the minimap2 chaining algorithm here.
+
+\paragraph*{Minimap2-like chaining}
+Formally, an \emph{anchor} is a 3-tuple $(x,y,w)$, representing a closed
+interval $[x-w+1,x]$ on a segment in the reference graph matching an interval
+$[y-w+1,y]$ on the query. Given a list of anchors sorted by $x$, let $f(i)$ be
+the maximal chaining score up to the $i$-th anchor in the list. $f(i)$ can be
+computed by:
+\begin{equation}\label{eq:dp}
+f(i)=\max\big\{\max_{i>j\ge1}\{f(j)+\alpha(j,i)-\beta(j,i)\},w_i\big\}
+\end{equation}
+where $\alpha(j,i)=\min\big\{\min\{y_i-y_j,x_i-x_j\},w_i\big\}$ is
+the number of matching bases between anchor $i$ and $j$.
+$\beta(j,i)$ is the gap penalty. Let $g_{ji}=|(y_i-y_j)-(x_i-x_j)|$
+be the gap length and $d_{ji}=\min\{y_i-y_j,x_i-x_j\}$ be the smaller distance
+between the two anchors.  Minigraph uses the following gap cost:
+$$
+\beta(j,i)=\left\{\begin{array}{ll}
+\infty & (g_{ji}>G) \\
+c_1\cdot g_{ji} + c_2\cdot d_{ji} + \log_2{g_{ji}} & (0<g_{ji}\le G) \\
+0 & (g_{ji}=0)\\
+\end{array}\right.
+$$
+where $G=100000$ in the graph construction mode, $c_1=e^{-dw}$ and
+$c_2=0.05\cdot e^{-dw}$. By default, $d=0.01$ is the expected per-base sequence
+divergence and $w=19$ is the minimizer length. In comparison, minimap2 applies
+$G=5000$, $c_1=0.19$ and $c_2=0$. Minigraph allows much larger gaps between
+minimizers and more heavily penalizes gaps.
+
+Solving Eq.~\ref{eq:dp} leads to an $O(n^2)$ algorithm where $n$ is the number
+of anchors. This algorithm is slow for large $n$. Minimap2 introduces
+heuristics to speed up the computation by approximating this equation. It works
+well for minimap2 that only allows small gaps and has base-level alignment as a
+fix to chaining errors. However, as minigraph intends to chain much longer
+gaps, the minimap2 algorithm occasionally misses the optimal alignment in long
+segmental duplications and produces false variations. Minigraph introduces a
+new heuristic to speed up chaining.
+
+\begin{figure}[tb]
+\centering
+\includegraphics[width=.36\textwidth]{Fig5}
+\caption{\csentence{Implementing 1-dimension Range-Min-Query (RMQ).} Given a
+  set of 2-tuples, a binary search tree is built for the first values in the
+  tuples. Each node $p$ in the tree is associated with a pointer. The pointer
+  points to the node that is in the subtree descended from $p$ and has the
+  minimal second value. In this example, ${\rm RMQ}(20,50)=14$.}\label{fig:rmq}
+\end{figure}
+
+\paragraph*{Dynamic 1-dimension Range-Min-Query}
+Before we move onto the minigraph solution, we will first introduce
+Range-Min-Query (RMQ). Given a set of 2-tuples $\{(y_i,s_i)\}$, ${\rm
+RMQ}(a,b)$ returns the minimum $s_j$ among $\{s_j:a\le y_j\le b\}$.
+We implemented 1-dimension RMQ with a modified AVL tree, a type of balanced
+binary search tree (Fig.~\ref{fig:rmq}). When performing ${\rm RMQ}(a,b)$, 
+we first find the smallest and the largest nodes within interval $[a,b]$ using
+the standard algorithm. In this example, the two nodes are (21,32) and (45,21),
+respectively. We then traverse the path between the two nodes to find the
+minimum. With a balanced tree structure, we do not need to descend into
+subtrees. The time complexity is $O(m\log m)$, where $m$ is the number of nodes
+in the tree. We can insert nodes to or delete nodes from the tree while
+maintaining the property of the tree. This achieves dynamic RMQ.
+
+\paragraph*{Chaining with a linear gap cost function}
+A linear gap cost takes the form of
+$\beta'(j,i)=c_1[(y_i-y_j)+(x_i-x_j)]$. Given a list of anchors
+$(x_i,y_i,w_i)$ sorted by position $x_i$, let
+\begin{equation}\label{eq:dp2}
+f'(i)=\max_{\substack{\text{$i>j\ge1$}\\ \text{$x_i-G\le x_j\le x_i-w_i$}\\ \text{$y_i-G\le y_j\le y_i-w_i$}}}\big\{f'(j)+w_j-\beta'(j,i)\big\}
+\end{equation}
+We can find the optimal $f'(i)$ in $O(n\log n)$ time with
+RMQ~\cite{DBLP:conf/wabi/AbouelhodaO03,Otto:2011aa}. To see that, define
+$$h'(j)=f'(j)+w_j+c_1(y_j+x_j)$$
+The following condition
+$$f'(j)+w_j-\beta'(j,i)>f'(k)+w_k-\beta'(k,i)$$
+is equivalent to $h'(j)>h'(k)$, independent of $i$. If we maintain ${\rm
+RMQ}_i$ as the binary tree that keeps $\{(y_j,-h'(j)):j<i,x_i-G\le x_j\le x_i-w_i\}$, we have
+$$
+f'(i)=-{\rm RMQ}_i(y_i-G,y_i-w_i)-c_1(x_i+y_i)
+$$
+This solves Eq.~\ref{eq:dp2} in $O(n\log n)$ time.
+
+\paragraph*{Minigraph linear chaining}
+While chaining with a linear gap cost function can be solved efficiently, we
+prefer more realistic cost function used in minimap2. In practical
+implementation, when we come to anchor $i$, we find the optimal predecessor $j_*$
+under the desired gap cost $\beta(j,i)$ for anchors $\{j:j<i,x_i-G'\le
+x_j<x_i,y_i-G'\le y_j<y_i\}$, where $G'<G$ is set to 10000 by default.
+Meanwhile, we use the RMQ-based algorithm to identify the anchor $j'_{*}$ optimal
+under the linear gap cost $\beta'(j,i)$. We choose $j'_*$ as the optimal
+predecessor if
+$$
+f(j_*)+\alpha(j_*,i)-\beta(j_*,i)<f(j'_*)+\alpha(j'_*,i)-\beta(j'_*,i)
+$$
+This may occasionally happen around long segmental duplications when the
+minimap2 heuristic misses the optimal solution. Effectively, minigraph does
+thorough search in a small window and approximate search in a large window
+using a faster but less sophisticated gap cost function.
+
+\subsubsection*{Graph chaining}
+
+Minigraph generates a set of linear chains $\{L_i\}$ with the procedure above
+that completely ignores the graph topology. It then applies another round of
+chaining taking the account of the topology.
+
+We say linear chain $L_i$ \emph{precedes} $L_j$, written as $L_i\prec L_j$, if
+(1) the ending coordinate of $L_i$ on the query sequence is smaller than the
+ending coordinate of $L_j$, and (2) there is a walk from $L_i$ to $L_j$ in the
+graph. If there are multiple walks from $L_i$ to $L_j$, minigraph enumerates
+the shortest 16 walks and chooses the walk with its length being the closest to
+the query distance between $L_i$ and $L_j$.
+
+Given a list of linear chains sorted by their ending coordinates on the query
+sequence, let $g(i)$ be the optimal graph chaining score up to linear chain
+$L_i$. We can compute $g(i)$ with another dynamic programming:
+$$
+g(i)=\max\big\{\max_{L_j\prec L_i}\{g(j)+\omega(L_j)-\beta(j,i)\},\omega(L_i)\big\}
+$$
+where $\beta(j,i)$ is the weight between $L_i$ and $L_j$. As minigraph does not
+perform base-level alignment, $\beta(j,i)$ is the same as the gap penalty
+function used for linear chaining. $\omega(L_i)$ is the optimal score of $L_i$
+computed during linear chaining.
+
+The procedure above has two limitations. First, when computing the weight
+between $L_i$ and $L_j$, minigraph largely ignores base sequences and only considers
+the distance between them on both the query and the graph. When there are
+multiple walks of similar lengths between $L_i$ and $L_j$, minigraph miss the
+graph chain that leads to the best base alignment. Although we added a
+heuristic by considering 17-mer matches between the query and the graph paths,
+we found this heursitc is not reliable in complex regions. Second, minigraph only
+enumerates the shortest 16 walks. In complex subgraphs, the optimal walk from
+$L_i$ to $L_j$ may not be among them. We plan to implement base
+alignment to address the limitations. We may use the current minigraph algorithm
+for easy cases and apply the more expensive base alignment when the current
+algorithm potentially fails.
+
+The graph chaining algorithm results in one or multiple graph chains.  A
+\emph{graph chain} is a list of anchors $(s_i,x_i,y_i,w_i)$, where
+$[x_i-w_i+1,x_i]$ on segment $s_i$ in the graph matches $[y_i-w_i+1,y_i]$ on
+the query sequence. A graph chain satisfies the following conditions: if $i<j$,
+$y_i<y_j$; if $i<j$ and $s_i=s_j$, we have $x_i<x_j$; if $s_i\not=s_{i+1}$, the
+two segments are adjacent on the graph. It is an extension to linear chains.
+
+\subsection*{The minigraph graph generation algorithm}
+
+Using the minimap2 algorithm~\cite{Li:2018ab}, minigraph identifies a set of
+\emph{primary chains} that do not greatly overlap with each other on the query
+sequence. A region on the query is considered to be \emph{orthogonal} to the
+reference if the region is contained in a primary chain longer than 100kb and
+it is not intersecting other primary chains longer than 20kb.
+
+Minigraph scans primary chains in orthogonal regions and identifies subregions
+where the query subsequences significantly differs from the corresponding
+reference subsequences. To achieve that, minigraph computes a score $h_i$ for
+each adjacent pair of anchors $(s_i,x_i,y_i,w_i)$ and
+$(s_{i+1},x_{i+1},y_{i+1},w_{i+1})$. Let $d^x_i$ be the distance between the
+two anchors on the graph and $d^y_i=y_{i+1}-y_i$ be the distance on the query
+sequence. $h_i$ is computed as
+\begin{equation}\label{eq:hi}
+h_i=\left\{\begin{array}{ll}
+-10 & \mbox{if $d^x_i=d^y_i\le w_{i+1}$} \\
+\eta\cdot\max\{d^x_i,d^y_i\} & \mbox{otherwise}\\
+\end{array}\right.
+\end{equation}
+where $\eta$ is the density of anchors averaged across all primary graph
+chains. Define $H(i,j)=\sum_{k=i}^j h_k$. A highly divergent region between the
+query and the graph will be associated with a large $H(i,j)$. Minigraph uses
+the Ruzzo-Tompa algorithm~\cite{DBLP:conf/ismb/RuzzoT99} to identify all
+maximal scoring intervals on list $(h_i)$, which correspond to divergent
+regions. In each identified divergent region, minigraph performs base
+alignment~\cite{Suzuki:2018aa,Li:2018ab} between the query and the graph
+sequences and retains a region if it involves an INDEL $\ge$100bp in length or
+a $\ge$100bp region with base-level identity below 80\%. In Eq.~\ref{eq:hi},
+-10 is an insensitive parameter due to the downstream filtering. In the end,
+minigraph augments the existing graph with identified variations
+(Fig.~\ref{fig:mg}b).
+
+\subsection*{Annotating variations}
+
+We applied RepeatMasker~\cite{Tarailo-Graovac:2009aa} v1.332 to classify
+interspersed repeats in the longest allele sequence of each variation.
+RepeatMasker is unable to annotate VNTRs with long motifs. It also often
+interprets VNTRs as impure STRs. Therefore, we did not use the RepeatMasker
+VNTR or STR annotations directly. Instead, we combined RepeatMasker and
+SDUST~\cite{Morgulis:2006aa} results to collect low-complexity regions (LCRs).
+We identified pure tandem repeats composed of a motif occurring twice or more
+(implemented in
+\href{https://github.com/lh3/etrf}{https://github.com/lh3/etrf}). An LCR is
+classified as VNTR if 70\% of the LCR is VNTR; similarly, an LCR is classified
+as STR if 70\% is STR; the rest are classified as ``Other-LCR'' in
+Fig.~\ref{fig:anno}. The annotation script is available in the minigraph GitHub
+repository.
+
+\subsection*{Creating blacklist regions}
+
+For each variation in the graph, we extend its genomic interval on GRCh38 by
+50bp from each end. We name this set of intervals as $I_0$. We align sequences
+inserted to GRCh38 against GRCh38 with ``minimap2 -cxasm20 -r2k'' and filter
+out alignments with mapping quality below 5. Let $I(a,b)$ be the set of GRCh38
+intervals that are contained in alignments with identity between $a$ and $b$.
+The blacklist regions are computed by $I_0\cup I(0,0.99)\setminus I(0.998,1)$,
+where ``$\cup$'' denotes the interval union operation and ``$\setminus$''
+denotes interval subtraction.
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%                                          %%
+%% Backmatter begins here                   %%
+%%                                          %%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+\begin{backmatter}
+
+\section*{Competing interests}
+The authors declare that they have no competing interests.
+
+\section*{Ethical Approval}
+Ethical approval was not needed for this study.
+
+\section*{Author's contributions}
+HL conceived the project, developed minigraph and drafted the manuscript.
+XF did the pseudogene analysis. CC helped with RepeatMasker annotation.
+All authors helped to revise the manuscript.
+
+\section*{Acknowledgements}
+We are grateful to Benedict Paten and Erik Garrison for discussions on
+pangenome graphs. We thank minigraph users who have suggested features and
+helped to fix various issues.
+
+\section*{Funding}
+This work is supported by National Institutes of Health (NIH) grant
+U01HG010961 and R01HG010040.
+
+\section*{Availability of data and materials}
+Minigraph is openly available at
+\href{https://github.com/lh3/minigraph}{https://github.com/lh3/minigraph}.
+This repository also includes the script to convert from the segment coordinate
+to the stable coordinate, to annotate variations and to generate blacklist
+regions from the graph. The companion gfatools is available at
+\href{https://github.com/lh3/gfatols}{https://github.com/lh3/gfatools}. The
+human and the great ape graphs are hosted at
+\href{ftp://ftp.dfci.harvard.edu/pub/hli/minigraph/}{ftp://ftp.dfci.harvard.edu/pub/hli/minigraph/}.
+The NA12878, NA24385 and PGP1 phased assemblies were downloaded from
+\href{ftp://ftp.dfci.harvard.edu/pub/hli/whdenovo/}{ftp://ftp.dfci.harvard.edu/pub/hli/whdenovo/}.
+Assemblies generated by McDonnell Genome Institute include 
+GCA\_001524155.4 for NA19240, GCA\_002180035.3 for HG00514, GCA\_002209525.2
+for HG01352, GCA\_002872155.1 for NA19434, GCA\_003574075.1 for HG02818,
+GCA\_003086635.1 for HG03486, GCA\_003086635.1 for HG03486, GCA\_003601015.1
+for HG03807, GCA\_002208065.1 for HG00733, GCA\_003070785.1 for HG02059,
+GCA\_008065235.1 for HG00268 and GCA\_007821485.1 for HG04217. Other assemblies
+are available from GenBank under accession GCA\_001297185.1 for
+CHM1~\cite{Huddleston:2017aa}, GCA\_000983455.1 for
+CHM13~\cite{Huddleston:2017aa}, GCA\_001750385.1 for AK1~\cite{Seo:2016aa},
+GCA\_002880755.3 for chimpanzee Clint~\cite{Kronenberg:2018aa},
+GCA\_900006655.3 for gorilla Susie~\cite{Gordon:2016kq}, GCA\_008122165.1 for
+gorilla Kamilah~\cite{Kronenberg:2018aa} and GCA\_002880775.3 for orangutan
+Susie~\cite{Kronenberg:2018aa}.
+
+
+\bibliographystyle{bmc-mathphys}
+\bibliography{minigraph}
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%                               %%
+%% Figures                       %%
+%%                               %%
+%% NB: this is for captions and  %%
+%% Titles. All graphics must be  %%
+%% submitted separately and NOT  %%
+%% included in the Tex document  %%
+%%                               %%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+%\section*{Figures}
+
+%\begin{figure}[h!]
+%  \caption{\csentence{Sample figure title.}
+%      Figure legend text.}
+%      \end{figure}
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%                               %%
+%% Tables                        %%
+%%                               %%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+%\section*{Tables}
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%                               %%
+%% Additional Files              %%
+%%                               %%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+%\section*{Additional Files}
+%  \subsection*{Additional file 1 --- Sample additional file title}
+%    Additional file descriptions text (including details of how to
+%    view the file, if it is in a non-standard format or the file extension).  This might
+%    refer to a multi-page table or a figure.
+
+%  \subsection*{Additional file 2 --- Sample additional file title}
+%    Additional file descriptions text.
+
+
+\end{backmatter}
+\end{document}