/td>

Mutation T@ster

statistics

March 4th, 2013

integration of the public version of HGMD

We have integrated the public version of the Human Gene Mutation Database (HGMD). The data includes the positions of the disease mutations and their HGMD ID. The disease alleles are not included so we cannot use HGMD for automatic predictions.
Whenever an HGMD public disease mutation is found at the same position as a variant, this will be written in the summary. We also place a direct hyperlink to the mutation in HGMD into the 'dbSNP / TGP / HGMD(public) / ClinVar' field, so you can check whether the HGMD mutation has the same allele as your variant (and whether the disease matches). Please note that you must be logged in at the HGMD site to make the hyperlink work - access to the public version is free but requires registration.

BLAST results of protein conservation analysis cached

We are happy to inform you that we were able to substantially improve MutationTaster's speed by caching the BLAST results which are generated for conservation analysis on protein level. Once a gene is queried via MutationTaster, the alignment of the wildtype protein sequence with the homolog sequences of up to ten other species will be stored in our database. The next time the same gene is queried, it is not necessary to invoke BLASTp again, but the alignments can be retrieved from our database, which is much faster than calling BLASTp.

new method for comparison of amino acid sequences

Formerly, we have used a Needleman-Wunsch application in order to compare the wildtype amino acid sequence with the mutated amino acid sequence and thus identify amino acid changes caused by a DNA mutation. Unfortunately, the Needleman-Wunsch application led to high CPU loads making our server suffer. For this reason, we decided to build a customized subroutine (basically by doing string diff) in order to identifychanges in the protein sequence. The waiving of the Needleman-Wunsch application further enhances MutationTaster's speed.

conservation analysis on nucleotide level is now optional

By default, nucleotide alignment is not run, since the BLAST call slows down MutationTaster and the results are not used by the Bayes Classifier anyway. You can ask MutatioTaster to assess conservation status on nucleotide level nevertheless by checking this option on the start page. As a (better) alternative to conservation analysis on nucleotide level, since the release of MutatiopnTaster2, we offer phyloP and PhastCons scores as a measure of evolutionary conservation (see documentation for further information on phyloP and PhastCons usage in MutationTaster).

February 7th, 2013

Grantham scores changed

For historical reasons, formerly, we have not displayed the original Grantham scores but scaled them down to values ranging from 0.0 to 6.0. Since the Grantham scores are not used by the Classifier to generate the prediction, but are only displayed for user information, we decided to switch to the original Grantham scores. These range from 0 to 215.

December 2012

MutationTaster 2

We are very happy to inform you about the availability of a new version of MutationTaster. It is not only based on the current version of Ensembl (v69), but also comes along with the following improvements:

usage of a new Bayes Classifier trained with disease-causing mutations from HGMD professional and with harmless polymorphisms from the 1000Genomes Project with better predictive performance
phyloP and phastCons values as a measure of evolutionary conservation
integration of known disease variants from dbSNP ClinVar
integration of regulatory features from Ensembl Regulation
possibility to query variants spanning an intron-exon boundary
possibility to turn off multi-species alignment on nucleotide level (improves speed)
revised display of results featuring the variation in HGVS nomenclature

More details can be found in the documentation.

query chromosomal positions

Upon great demand, we built an extra interface to query MutationTaster with chromosomal positions. MutationTaster analyses the submitted variant at the given position in all feasible Ensembl transcripts and puts out a table with summarized results as well as the traditional detailed results.

QueryEngine replaces BatchQueries and NGSpipeline

While a few years ago, Sanger sequencing produced only a handful of variants at once, people nowadays run more and more Next Generation Sequencing projects. We felt that our services BatchQueries and NGSpipeline doesn't meet the requirements for data analysis of these high-throughput techniques any more and thus decided to build a much faster and better QueryEngine. It enables you to analyse the results of Next Generation Sequencing (e.g. Whole Exome Sequencing) projects in a comfortable way and reasonable time. The input format for the QueryEngine is VCF, which is the output format of many variant calling algorithms and emerged to be the standard format in the field. The QueryEngine provided here runs stable, but a beta version exists and will soon be released with additional features. Look up more details in the QueryEngine documentation. With the release of the QueryEngine we will end our support for the deprecated BatchQueries and NGSpipeline. The scripts can still be downloaded here (BatchQueries | NGSpipeline), but are not further maintained and supported. We friendly request your sympathy for this decision. It was mainly made because we are only two people and for time issues cannot handle the support questions concerning the old services any more.

Contact

In case you discover bugs, have suggestions or questions, please write an e-mail to
Jana Marie Schwarz (jana-marie.schwarz AT charite.de) or to
Dominik Seelow (dominik.seelow AT charite.de).
We also appreciate hearing about your general experiences using MutationTaster.

Comparison of different predictive tools

MutationTaster2 - statistics

This is somewhat a placeholder... I will add more data soon...

cross-validation

combined statistics

model	parameter	value	standard deviation
simple_aae	NPV	0.877	0.008
simple_aae	PPV	0.895	0.005
simple_aae	accuracy	0.886	0.004
simple_aae	accuracy_DM	0.895	0.005
simple_aae	accuracy_SNP	0.877	0.008
simple_aae	n	4000.000	0.000
simple_aae	sensitivity	0.879	0.007
simple_aae	specifity	0.893	0.004
without_aae	NPV	0.957	0.003
without_aae	PPV	0.888	0.006
without_aae	accuracy	0.922	0.004
without_aae	accuracy_DM	0.888	0.006
without_aae	accuracy_SNP	0.957	0.004
without_aae	n	4000.000	0.000
without_aae	sensitivity	0.954	0.004
without_aae	specifity	0.895	0.005
complex_aae	NPV	0.869	0.032
complex_aae	PPV	0.944	0.004
complex_aae	accuracy	0.907	0.017
complex_aae	accuracy_DM	0.944	0.004
complex_aae	accuracy_SNP	0.869	0.032
complex_aae	n	400.000	0.000
complex_aae	sensitivity	0.879	0.026
complex_aae	specifity	0.939	0.005


#model	test	values
simple_aae	NPV	0.872; 0.881; 0.889; 0.8695; 0.8725
simple_aae	PPV	0.8885; 0.896; 0.891; 0.8985; 0.9005
simple_aae	accuracy	0.88; 0.888; 0.89; 0.884; 0.886
simple_aae	accuracy_DM	0.888; 0.896; 0.891; 0.898; 0.9
simple_aae	accuracy_SNP	0.872; 0.881; 0.889; 0.87; 0.872
simple_aae	n	4000; 4000; 4000; 4000; 4000
simple_aae	sensitivity	0.874077717658633; 0.882758620689655; 0.889221556886228; 0.873177842565598; 0.875972762645914
simple_aae	specifity	0.886629384850025; 0.894416243654822; 0.890781563126252; 0.895468589083419; 0.897633744855967
without_aae	NPV	0.951; 0.9565; 0.9585; 0.9595; 0.9585
without_aae	PPV	0.883; 0.8925; 0.894; 0.881; 0.89
without_aae	accuracy	0.917; 0.925; 0.926; 0.92; 0.924
without_aae	accuracy_DM	0.883; 0.892; 0.894; 0.881; 0.89
without_aae	accuracy_SNP	0.951; 0.957; 0.959; 0.96; 0.959
without_aae	n	4000; 4000; 4000; 4000; 4000
without_aae	sensitivity	0.947424892703863; 0.953525641025641; 0.955638695884554; 0.95604991861096; 0.955448201825013
without_aae	specifity	0.890449438202247; 0.898966165413534; 0.900422733677783; 0.889661566991191; 0.897051941974731
complex_aae	NPV	0.9; 0.83; 0.89; 0.885; 0.84
complex_aae	PPV	0.95; 0.945; 0.94; 0.945; 0.94
complex_aae	accuracy	0.925; 0.888; 0.915; 0.915; 0.89
complex_aae	accuracy_DM	0.95; 0.945; 0.94; 0.945; 0.94
complex_aae	accuracy_SNP	0.9; 0.83; 0.89; 0.885; 0.84
complex_aae	n	400; 400; 400; 400; 400
complex_aae	sensitivity	0.904761904761905; 0.847533632286996; 0.895238095238095; 0.891509433962264; 0.854545454545454
complex_aae	specifity	0.947368421052632; 0.937853107344633; 0.936842105263158; 0.941489361702128; 0.933333333333333