Skip to content
Snippets Groups Projects
Commit 6e5795bc authored by Laros's avatar Laros
Browse files

Added a first draft of the Mutalyzer 2.0 paper. 


git-svn-id: https://humgenprojects.lumc.nl/svn/mutalyzer/trunk@622 eb6bd6ab-9ccd-42b9-aceb-e2899b4a52f1
parent a7248a86
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
doc/Mutalyzer 2.0/Makefile 0 → 120000
+ 1
− 0
View file @ 6e5795bc
/home/jfjlaros/projects/skel/Makefile
\ No newline at end of file
doc/Mutalyzer 2.0/paper.bbl 0 → 100644
+ 23
− 0
View file @ 6e5795bc
\begin{thebibliography}{1}
\bibitem{HGVS}
{Human Genome Variation Society}.
\newblock \begin{small}\texttt{http://www.hgvs.org/mutnomen/}\end{small}.
\bibitem{hgvs_bnf}
J.F.J. Laros, A.~Blavier, J.T. den Dunnen, and P.E.M. Taschner.
\newblock A formalized description of the standard human variant nomenclature
in extended backus-naur form.
\newblock {\em BMC Bioinformatics}, 12(Suppl 4):S5, 2011.
\bibitem{NCBI}
{National Center for Biotechnology Information}.
\newblock \begin{small}\texttt{http://www.ncbi.nlm.nih.gov/}\end{small}.
\bibitem{Mutalyzer}
M.~Wildeman, E.~van Ophuizen, J.~T. den Dunnen, and P.~E. Taschner.
\newblock Improving sequence variant descriptions in mutation databases and
literature using the {Mutalyzer} sequence variation nomenclature checker.
\newblock {\em Human Mutation}, 29:6–--13, 2008.
\end{thebibliography}
doc/Mutalyzer 2.0/paper.tex 0 → 100644
+ 330
− 0
View file @ 6e5795bc
\documentclass{article}
\usepackage{ltxtable}
\frenchspacing
\title{\Huge Mutalyzer 2.0}
\author{Jeroen F.J. Laros, Martijn Vermaat, Gerben R. Stouten,\\
Johan T. den Dunnen, Peter E.M. Taschner
\vspace{10pt}\\
Department of Human Genetics\\
Center for Human and Clinical Genetics\\
\texttt{j.f.j.laros@lumc.nl}}
\date{\today}
%\setlength{\parindent}{0pt}
\begin{document}
\maketitle
\begin{abstract} \noindent
\end{abstract}
\section{Introduction}\label{introduction}
\cite{Mutalyzer}
\cite{HGVS}
Currently GenBank and LRG reference supported.
Any organism, including translation.
Phased variants, of all genes containing variants, half of them have more than
one.
\section{Background}\label{background}
\section{Materials and Methods}
%Distributed setup, multiple servers can be used, cache and database are
%synchronised.
The suite is logically divided in modules which can be combined to make several
interfaces.
\begin{table}[]
\begin{center}
\caption{Modules.}
\begin{tabular}{l|l}
name & function\\
\hline
HGVS parser & Check the HGVS nomenclature.\\
Retriever & Retrieve a reference sequence.\\
Crossmapper & Convert positions from \texttt{g.} to \texttt{n.} or
\texttt{c.} and vice versa.\\
Database & Interface to Mutalyzer's internal database.\\
Mutator & Apply variants to a reference sequence.\\
GenRecord & Generalisation of reference sequences.\\
VariantChecker & Semantic checks.
\end{tabular}
\end{center}
\label{tab:modules}
\end{table}
In Table~\ref{tab:modules} we see a list of core modules in the Mutalyzer
suite, the functionality of each module is described in the following sections.
Database:
\begin{itemize}
\item Mapping of transcripts and genes.
\item Cache administration.
\item Batch checker.
\end{itemize}
\subsection{Name Checker} \label{subsec:namecheck}
The formalisation of the HGVS nomenclature~\cite{hgvs_bnf}, made it possible to
implement a context free grammar parser that encompasses the complete
nomenclature, including nesting and other newly added features. Although the
semantic checks have not been implemented for the complete nomenclature, we
have implemented the recognition of allele descriptions. These descriptions
are, apart from simple variants consisting of one change, the most occurring
ones. With allele descriptions we can describe a large number of changes, i.e.,
all descriptions that need no information from outside the reference sequence.
We define a \emph{raw variant} as an elementary variant, e.g., a substitution,
deletion, insertion, etc. An \emph{allele description} is obtained by
concatenation of raw variants. Consider the following description
\texttt{NM\_002001.2:c.[10del;22C>T;101\_119inv]}. The part between brackets is
the allele description, which consists of raw variants separated by a semicolon
(\texttt{10del}, \texttt{22C>T} and \texttt{101\_119inv}).
After the description is parsed, the \emph{reference sequence}, which is part
of the description, is retrieved from a reference sequence repository (e.g.,
the NCBI~\cite{NCBI} or the EBI). The reference sequence (in either GenBank or
LRG format) is then parsed and any missing data is retrieved to prepare for the
next step. See Section~\ref{sec:enrichment} for a complete description.
With the parsed variant description and the information from the reference sequence, the variant can be simulated to produce the observed
sequence. In this simulation all raw variants are visualised and checked. The
checks on the raw variants is extensive.
First, we check whether the minimal description is used. In case of a
\texttt{delins}, one can for example add reference bases to the inserted
sequence, adding nothing to the description. If, for example, the reference
sequence is \texttt{AACGTAA}, we can define the following deletion-insertion:
\texttt{3\_4delinsTT}, resulting in an observed sequence of \texttt{AATTTAA}.
The same result will be obtained if we define the variant as
\texttt{2\_6delinsATTTA}. This latter description can be minimised by
calculating and removing the \emph{longest common prefix} and the \emph{longest
common suffix} of the deleted and the inserted sequence.
In case of an inversion, a prefix of the inversion can be the reverse
complement of a suffix, we call this a \emph{partial palindrome}. The
description of an inversion is minimised in a similar way as described above.
\begin{table}
\begin{center}
\caption{Disambiguation of raw variant types.}
\begin{tabular}{l|l}
type & simplification\\
\hline
\texttt{delins} & \texttt{del}, \texttt{ins}, \texttt{subst},
\texttt{inv}, \texttt{dup}\\
\texttt{ins} & \texttt{dup}\\
\texttt{inv} & \texttt{subst}
\end{tabular}
\end{center}
\label{tab:typedisambiguation}
\end{table}
After the minimisation step, a disambiguation scheme is used to check whether
the type of the raw variant is correct. In Table~\ref{tab:typedisambiguation}
we see which variant types can possibly be written as a simpler type.
Finally, a deletion or insertion is shifted to the most 3' position possible.
We use a \emph{rolling} algorithm that takes circular permutations of the
deleted or inserted sequence into account. If for example, we insert the
sequence \texttt{TCCA} in a reference sequence \texttt{CATC}, the
algorithm will correct the description \texttt{2\_3insTCCA} to
\texttt{3\_4insCCAT}. Note that this method works for both the forward as well
as the reverse strand. If a gene resides on the reverse strand, the position
will be shifted in the opposite direction of that of the genomic one.
Furthermore, if an insertion or a deletion is described on a transcript, the
position will not be shifted over a splice site.
%Optional arguments (\texttt{10\_12delAAT}) are checked.
After the simulation of the variant, we have the observed sequence. We use this
observed sequence to do basic effect prediction.
For all the annotated transcripts in the reference sequence, the corrected
variant description is shown, as well as a list of protein descriptions. Each
of the DNA variant descriptions can be selected for a more detailed analysis.
In the detailed analysis the reference protein and the variant protein is
visualised and the area of change is highlighted. For the selected transcript,
a list of exons start and end positions are given, as well as the CDS start and
end positions.
For all raw variants, effects on restriction sites are calculated. A table is
generated that contains the number of the raw variant, a list of removed
restriction sites and a list of added restriction sites.
Deletion of exons as well as partial exons (resulting in a fusion exon) is
supported.
Gives informative warnings when a variant is near a splice site.
%Warns when a change has no effect (\texttt{10A>A}, inversion of a palindrome,
%etc.).
Supports ``fuzzy'' positions.
\subsection{Syntax Checker}
The \emph{Syntax Checker} is an interface to the HGVS parser, described in
Section~\ref{subsec:namecheck}, only. This interface will only return whether
or not the syntax of a variant description is correct, no semantic check is
performed. If no reference sequence is available, one might be restricted to
checking the syntax only. An other reason for using the syntax checker is to
check large quantities of descriptions in a small amount of time. Since there
is no communication with reference sequence repositories, this check is
extremely quick.
\subsection{Position Converter}
The \emph{position converter} is an interface to the HGVS parser, the
crossmapper and the database. With this interface, we can convert a description
that uses a RefSeq transcript as reference sequence to a description on a
chromosomal reference sequence and vice versa. The mapping information is
retrieved daily from the NCBI. Currently Human genome build 18 and 19 are
supported.
This interface can be used to quickly convert variants found by a high
throughput screening, e.g., an NGS experiment. This enables people to annotate
their NGS experiments with informative HGVS descriptions. An other use of this
interface is to convert (or lift over) a description from one transcript to an
other, or to transcripts of other (overlapping) genes. Finally, by using
transcripts that are mapped to both hg18 as well as to hg19, we can convert a
chromosomal description from one build to an other. Potentially, we can even
lift over descriptions to other species.
\subsection{SNP Converter}
For converting a DbSNP~\cite{DBSNP} id to an HGVS description, the \emph{SNP
converter} can be used. This interface retrieves the annotated HGVS
descriptions from the NCBI.
\subsection{Name Generator}
Educational interface for those who are not familiar with the HGVS
nomenclature.
Constructed variant description can be checked (clickable) with the name
checker.
\subsection{Reference File Loader}
For reference sequences unknown to the NCBI or EBI, we have created the
\emph{reference file loader}.
\begin{enumerate}
\item Upload a local file. \label{item:local}
\item Download a reference sequence by supplying a URL.
\item Retrieve part of the reference genome for a (HGNC) gene symbol
\begin{itemize}
\item Most recent build is used for the organism.
\item The orientation of the slice is selected automatically.
\item Flanking ranges.
\end{itemize}
\item Retrieve a range of a chromosome by accession number
\begin{itemize}
\item Choose orientation.
\end{itemize}
\item Retrieve a range of a chromosome by name
\end{enumerate}
Reference sequences are stored in a cache.
MD5sum is used to identify the file.
\begin{itemize}
\item Prevents re-uploading, the same UD is returned.
\item Enables the retrieval of reference sequences after it has vanished from
the cache, except for~\ref{item:local}.
\end{itemize}
\subsection{Batch Jobs}
For the Name Checker, Syntax Checker, Position Converter and SNP Converter.
Formats (automatically detected):
\begin{itemize}
\item Tab delimited text file / CSV file
\item Microsoft Excel file
\item OpenOffice ODS file
\end{itemize}
Each row consists of one or more tab delimited fields, where every field
contains a single variant description (or dbSNP rs number in case of the SNP
Converter). Note that all rows must have the same number of fields.
For backwards compatibility, the format used by Mutalyzer~1.0.3 is also
accepted.
The output of a Mutalyzer Batch run is a tab delimited CSV file, which has a
header-row to clarify the results. We recommend opening the file in a
spreadsheet program, such as OpenOffice Calc or Microsoft Excel. Note that
empty lines are removed from the batch input file.
Batch jobs are interleaved, so that even if large jobs are submitted, small
jobs will still finish soon.
Scheduler can be stopped, will resume even after power failure.
\subsection{Webservices}
\begin{itemize}
\item SOAP
\item HTTP/RPC+JSON
\end{itemize}
Well documented API online.
Examples for other usage (textmining) given.
Someone made a java client? Perhaps add link?
\subsection{Feedback}
Trac system for requests, documentation and error reporting.
\subsection{Experimental description extractor}
Generates a description from two sequences.
\begin{itemize}
\item Use after applying a variant in the Name Checker.
\item Compare two reference sequences.
\end{itemize}
Will solve the combining and splitting of variants problem. True
disambiguation.
\section{LOVD~3.0}
Uses the Mutalyzer~2.0 Webservices for:
\begin{itemize}
\item Retrieving a reference sequence (add new gene).
\item Mapping descriptions.
\item Converting descriptions to other transcripts.
\item Checking variant descriptions.
\item \ldots
\end{itemize}
\section{Conclusions and further research}\label{conclusion}
EMBL reference sequences.
Description extractor in Name checker.
Nesting.
\bibliography{$HOME/projects/bibliography}{}
\bibliographystyle{plain}
\appendix
\section{Webservices}
\LTXtable{\textwidth}{webservices.tex}
\section{Annotation enrichment} \label{sec:enrichment}
The reference sequence annotation enrichment consists mainly of the linking of
annotated transcripts to their \emph{coding sequence} (CDS) and protein. In
many cases, (especially in GenBank files) there is no direct link between a
transcript and its CDS, making it impossible to reconstruct the layout of the
transcript and thereby the biological effect of a variant. Therefore we have
developed an extensive set of methods to accomplish this link. First we select
the CDSs that are consistent with a certain transcript, then we try to find a
connection between the CDS and the transcript by comparing the locus tags. If
the locus tag is not present, we try to retrieve the link between the
identifier of the transcript and the accession number of the protein from the
NCBI. If this is also not available, we use the product tag. The method used
for connecting the CDS and transcript is reported.
\end{document}
doc/Mutalyzer 2.0/webservices.tex 0 → 100644
+ 57
− 0
View file @ 6e5795bc
\begin{longtable}{l|X}
\caption{List of webservices.}\\
name & description\\
\hline
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
placeholder & \\
checkSyntax & Checks the syntax of a variant.\\
chromAccession & Get the accession number of a chromosome, given a
name.\\
chromosomeName & Get the name of a chromosome, given a chromosome
accession number.\\
getCache & Get a list of entries from the local cache created
since given date.\\
getGeneAndTranscript & \\
getGeneName & Find the gene name associated with a transcript.\\
getTranscripts & Get all the transcripts that overlap with a
chromosomal position.\\
getTranscriptsAndInfo & Given a genomic reference, return all its
transcripts with their transcription/cds start/end sites and exons.\\
getTranscriptsByGeneName & \\
getTranscriptsMapping & Get all the transcripts and their info that
overlap with a range on a chromosome.\\
getTranscriptsRange & Get all the transcripts that overlap with a range
on a chromosome.\\
getchromName & Get the chromosome name, given a transcript
identifier (NM number).\\
getdbSNPDescriptions & Lookup HGVS descriptions for a dbSNP rs
identifier.\\
info & Gives some static application information, such as
the current running version.\\
mappingInfo & Search for an NM number in the MySQL database, if
the version number matches, get the start and end positions in a variant
and translate these positions to \texttt{g.} notation if the variant is in
\texttt{c.} notation and vice versa.\\
numberConversion & Converts \texttt{c.} to \texttt{g.} notation or
vice versa.\\
ping & Simple function to test the interface.\\
runMutalyzer & Run the Mutalyzer name checker.\\
sliceChromosome & \\
sliceChromosomeByGene & \\
transcriptInfo & Search for an NM number in the MySQL database, if
the version number matches, the transcription start and end and CDS end in
\texttt{c.} notation is returned.\\
upLoadGenBankLocalFile & \\
upLoadGenBankRemoteFile & \\
\label{tab:webservices}
\end{longtable}
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment