Title: Statistical Analyses of De Novo Genetic Variants
Version: 0.2.0
Date: 2016-08-01
Description: An integrated toolset for the analysis of de novo (sporadic) genetic sequence variants. denovolyzeR implements a mutational model that estimates the probability of a de novo genetic variant arising in each human gene, from which one can infer the expected number of de novo variants in a given population size. Observed variant frequencies can then be compared against expectation in a Poisson framework. denovolyzeR provides a suite of functions to implement these analyses for the interpretation of de novo variation in human disease.
Depends: R (≥ 3.1.0)
Imports: dplyr (≥ 0.3), reshape2 (≥ 1.4)
License: GPL-3
LazyData: true
Suggests: knitr, rmarkdown
VignetteBuilder: knitr
URL: http://denovolyzeR.org
BugReports: http://github.com/jamesware/denovolyzeR/issues
RoxygenNote: 5.0.1
NeedsCompilation: no
Packaged: 2016-08-01 12:18:10 UTC; jamesware
Author: James Ware [aut, cre], Jason Homsy [ctb], Kaitlin Samocha [ctb]
Maintainer: James Ware <j.ware@imperial.ac.uk>
Repository: CRAN
Date/Publication: 2016-08-01 14:55:41

A package for the analysis of de novo sequencing variants

Description

A package for the analysis of de novo sequencing variants

Author(s)

James Ware j.ware@imperial.ac.uk

References

http://github.com/jamesware/denovolyzeR

Permutes x variants across a genelist, and counts genes with multiple hits

Description

An internal function called by denovolyzeMultiHits

Usage

PermuteMultiHits(x, y, nperms = 100, class = "lof", geneId = "geneName",
  includeGenes = "all", probTable = pDNM)

Arguments

x

Total number of de novo variants observed in dataset

y

Number of genes with >1 de novo variant (of class "class") in the population

nperms

Number permutations

class

In c("lof","mis","syn","prot")

geneId

Gene identifier used. One of "hgncID", "hgncSymbol", "enstID", "ensgID" or "geneName" (default, equals ensembl "external_gene_name")

includeGenes

Genes to include in analysis. "all" or a vector of gene names.

probTable

Probability table. A user-defined table of probabilities can be provided here, to replace the probability table included in the package.

Value

Returns a named vector of 5 values

See Also

denovolyzeMultiHits

de novo variants found in 1,078 autism trios

Description

de novo variants found in 1,078 autism trios, published in Nature Genetics(http://www.nature.com/doifinder/10.1038/ng.3050)

Format

A data frame with 1096 obs of 2 variables:

gene

Gene symbol of gene containing de novo variant

class

Functional class of variant: "syn" = synonymous, "mis" = missense, "non" = nonsense, "splice" = canonical splice site, "frameshift" = frameshift indel

References

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4222185/

Evaluates burden of de novo variation against expectation

Description

Determines whether the test population carry more de novo variants than expected. Variants may be grouped by variant class (e.g. are there more LOF variants than expected, across the whole dataset?), or by gene (are there more variants of a given class in SCN2A?).

Usage

denovolyze(genes, classes, nsamples, groupBy = "class",
  includeGenes = "all", includeClasses = c("syn", "mis", "misD", "non",
  "stoploss", "startgain", "splice", "frameshift", "lof", "prot", "protD",
  "all"), geneId = "geneName", signifP = 3, roundExpected = 1,
  probTable = NULL, misD = NULL)

denovolyzeByClass(genes, classes, nsamples, groupBy = "class",
  includeGenes = "all", includeClasses = c("syn", "mis", "lof", "prot",
  "all"), geneId = "geneName", signifP = 3, roundExpected = 1,
  probTable = NULL)

denovolyzeByGene(genes, classes, nsamples, groupBy = "gene",
  includeGenes = "all", includeClasses = c("lof", "prot"),
  geneId = "geneName", signifP = 3, roundExpected = 1, probTable = NULL)

Arguments

genes

A vector of genes containing de novo variants.

classes

A vector of classes of de novo variants. Standard supported classes are "syn" (synonymous), "mis" (missense), "non" (nonsense), "splice" (splice), "frameshift" (frameshift) and "lof" (loss of function = non + splice + frameshift). Additional classes that are supported by the code, but are not included in the built-in probability tables, are "stoploss","startloss", "misD" (damaging missense). These labels may be used for user-supplied probability tables. If "misD" is present, then "mis" (in the input) implies non-damaging missense.

nsamples

Number of individuals considered in de novo analysis.

groupBy

Results can be tabulated by "gene", or by variant "class"

includeGenes

Genes to include in analysis. "all" or a vector of gene names.

includeClasses

Determines which variant classes are tabulated in output. In addition to the input classes, summaries can be produced for "prot" (protein-altering = mis + lof), "all", and "protD" (protein damaging = misD + lof, only available if misD included in user-specified probability table). If "misD" is present, then "mis" will return statistics for all missense. Non-damaging missense are not analysed separately.

geneId

Gene identifier used. One of "hgncID", "hgncSymbol", "enstID", "ensgID" or "geneName" (default, equals ensembl "external_gene_name")

signifP

Number of significant figures used to round p-values in output.

roundExpected

Number of decimal places used to round expected burdens in output.

probTable

Probability table. A user-defined table of probabilities can be provided here, to replace the probability table included in the package.

misD

If the user-specified probability table contains probabilities for a sub-category of missense variants (e.g. predicted to be damaging by an in silico algorithm), this column should be called misD, or the alternative name should be specified here.

Details

Analyses can be restricted to a subset of genes, and/or a subset of variant classes

See vignette("denovolyzeR_intro") for more information.

Value

Returns a data frame

Functions

Examples


### denovolyze

denovolyze(genes=autismDeNovos$gene,
           classes=autismDeNovos$class,
           nsamples=1078)

### denovolyzeByClass

denovolyzeByClass(genes=autismDeNovos$gene,
                  classes=autismDeNovos$class,
                  nsamples=1078)

# this convenience function is identical to:

denovolyze(genes=autismDeNovos$gene,
           classes=autismDeNovos$class,
           nsamples=1078,
           groupBy="class",
           includeClasses=c("syn","mis","lof","prot","all"),
           includeGenes="all"
           )

### denovolyzeByGene

denovolyzeByGene(genes=autismDeNovos$gene,
                 classes=autismDeNovos$class,
                 nsamples=1078)

# this is identical to:

denovolyze(genes=autismDeNovos$gene,
           classes=autismDeNovos$class,
           nsamples=1078,
           groupBy="gene",
           includeClasses=c("lof","prot"),
           includeGenes="all"
           )

Determine significance of genes with multiple de novos

Description

Are there more genes containing >1 de novos than expected?

Usage

denovolyzeMultiHits(genes, classes, nsamples, nperms = 100,
  includeGenes = "all", includeClasses = c("syn", "mis", "lof", "prot",
  "all"), nVars = "actual", geneId = "geneName", probTable = NULL,
  misD = NULL, signifP = 3, roundExpected = 1)

Arguments

genes

A vector of genes containing de novo variants.

classes

A vector of classes of de novo variants. Standard supported classes are "syn" (synonymous), "mis" (missense), "non" (nonsense), "splice" (splice), "frameshift" (frameshift) and "lof" (loss of function = non + splice + frameshift). Additional classes that are supported by the code, but are not included in the built-in probability tables, are "stoploss","startloss", "misD" (damaging missense). These labels may be used for user-supplied probability tables. If "misD" is present, then "mis" (in the input) implies non-damaging missense.

nsamples

Number of individuals considered in de novo analysis.

nperms

Number of permutations

includeGenes

Genes to include in analysis. "all" or a vector of gene names.

includeClasses

Determines which variant classes are tabulated in output. In addition to the input classes, summaries can be produced for "prot" (protein-altering = mis + lof), "all", and "protD" (protein damaging = misD + lof, only available if misD included in user-specified probability table). If "misD" is present, then "mis" will return statistics for all missense. Non-damaging missense are not analysed separately.

nVars

Select whether expected number of multihits is determined by "expected" total number of variants , or "actual" total. Actual (default) is more conservative.

geneId

Gene identifier used. One of "hgncID", "hgncSymbol", "enstID", "ensgID" or "geneName" (default, equals ensembl "external_gene_name")

probTable

Probability table. A user-defined table of probabilities can be provided here, to replace the probability table included in the package.

misD

If the user-specified probability table contains probabilities for a sub-category of missense variants (e.g. predicted to be damaging by an in silico algorithm), this column should be called misD, or the alternative name should be specified here.

signifP

Number of significant figures used to round p-values in output.

roundExpected

Number of decimal places used to round expected burdens in output.

Details

See vignette (denovostats_intro) for more information.

Value

Returns a data.frame

Examples

denovolyzeMultiHits(genes=autismDeNovos$gene,
                    classes=autismDeNovos$class,
                    nsamples=1078)

FMRP genes

Description

837 genes found to interact with the fragile X mental retardation protein (FMRP)

Format

A vector of gene symbols

References

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4222185/

http://dx.doi.org/10.1016/j.cell.2011.06.013

Checks input for errors

Description

An internal function to check inputs

Usage

parseInput(genes = genes, classes = classes, nsamples = nsamples,
  groupBy = groupBy, includeGenes = includeGenes,
  includeClasses = includeClasses, geneId = geneId, signifP = signifP,
  roundExpected = roundExpected, probTable = NULL)

Arguments

genes

A vector of genes containing de novo variants.

classes

A vector of classes of de novo variants. Standard supported classes are "syn" (synonymous), "mis" (missense), "non" (nonsense), "splice" (splice), "frameshift" (frameshift) and "lof" (loss of function = non + splice + frameshift). Additional classes that are supported by the code, but are not included in the built-in probability tables, are "stoploss","startloss", "misD" (damaging missense). These labels may be used for user-supplied probability tables. If "misD" is present, then "mis" (in the input) implies non-damaging missense.

nsamples

Number of individuals considered in de novo analysis.

groupBy

Results can be tabulated by "gene", or by variant "class"

includeGenes

Genes to include in analysis. "all" or a vector of gene names.

includeClasses

Determines which variant classes are tabulated in output. In addition to the input classes, summaries can be produced for "prot" (protein-altering = mis + lof), "all", and "protD" (protein damaging = misD + lof, only available if misD included in user-specified probability table). If "misD" is present, then "mis" will return statistics for all missense. Non-damaging missense are not analysed separately.

geneId

Gene identifier used. One of "hgncID", "hgncSymbol", "enstID", "ensgID" or "geneName" (default, equals ensembl "external_gene_name")

signifP

Number of significant figures used to round p-values in output.

roundExpected

Number of decimal places used to round expected burdens in output.

probTable

Probability table. A user-defined table of probabilities can be provided here, to replace the probability table included in the package.

Value

warning or error if any invalid input, else assigns variables back to parent function

Displays underlying de novo probability tables

Description

Tabulates probability of de novo variant for each protein-coding variant class, for each gene. Values are probability of a de novo variant per chromosome per generation. i.e. expected number of de novos for a given gene/class = p * 2 * nsamples.

Usage

viewProbabilityTable(format = "wide")

Arguments

format

option to display table in wide format (default; one line per gene), or long format