Package 'rfPred'

Assign functional prediction rfPred scores to human missense variants (random forest method based on SIFT, Polyphen2, PhyloP, LRT and Mutation Taster)

Description

The package provides a function which returns the rfPred score for a list of non-synonymous missense variants. All the rfPred scores are pre-calculated and stored in a TabixFile available on a server and which can be downloaded for using the package while not connected on the Internet. The package does not work without an access to the TabixFile. However, a toy example on the chromosome Y is available within the package to test the rfPred_scores function. curves with numbers of subjects at risk, compare data sets, display spaghetti-plot, build multi-contingency tables...

Author(s)

Fabienne Jabot-Hanin, Hugo Varet and Jean-Philippe Jais

References

dbNSFP database: Liu X, Jian X and Boerwinkle E. 2011. dbNSFP: a lightweight database of human non-synonymous SNPs and their functional predictions. Human Mutation. 32:894-899.

rfPred method: Jabot-Hanin F, Varet H, Tores F and Jais J-P. 2013. rfPred: a new meta-score for functional prediction of missense variants in human exome (submitted).

---

Toy example of GRanges object

Description

Toy example of GRanges object

Format

A GRanges object with 11 rows and several columns:

seqnames

Chromosome number (only Y in this example)

ranges

IRanges object for which start=end: position on the chromosome

reference

Referent nucleotid (A, C, G or T)

alteration

Alteration nucleotid (A, C, G or T)

---

Assign functional prediction rfPred scores to human missense variants

Description

rfPred is a statistical method which combines 5 algorithms predictions in a random forest model: SIFT, Polyphen2, LRT, PhyloP and MutationTaster. These scores are available in the dbNFSP database for all the possible missense variants in hg19 version, and the package rfPred gives a composite score more reliable than each of the isolated algorithms.

Usage

rfPred_scores(
  variant_list,
  data = system.file("extdata/chrY_rfPred.txtz", package = "rfPred"),
  index = system.file("extdata/chrY_rfPred.txtz.tbi", package = "rfPred"),
  all.col = FALSE,
  file.export = NULL,
  n.cores = 1
)

rfPred_scores(variant_list,
                     data=system.file("extdata/chrY_rfPred.txtz", package="rfPred"),
                     index=system.file("extdata/chrY_rfPred.txtz.tbi", package="rfPred"),
                     all.col=FALSE, file.export=NULL, n.cores=1)

## S4 method for signature 'character'
rfPred_scores(
  variant_list,
  data = system.file("extdata/chrY_rfPred.txtz", package = "rfPred"),
  index = system.file("extdata/chrY_rfPred.txtz.tbi", package = "rfPred"),
  all.col = FALSE,
  file.export = NULL,
  n.cores = 1
)

## S4 method for signature 'GRanges'
rfPred_scores(
  variant_list,
  data = system.file("extdata/chrY_rfPred.txtz", package = "rfPred"),
  index = system.file("extdata/chrY_rfPred.txtz.tbi", package = "rfPred"),
  all.col = FALSE,
  file.export = NULL,
  n.cores = 1
)

Arguments

variant_list	A variants list in a data.frame containing 4 or 5 columns: chromosome number, hg19 genomic position on the chromosome, reference nucleotid, variant nucleotid and uniprot protein identifier (optional); or a character string of the path to a VCF (Variant Call Format) file; or a GRanges object with metadata containing textually reference, alteration and proteine (optional) columns names for reference and alteration
data	Path to the compressed TabixFile, either on the server (default) or on the user's computer
index	Path to the index of the TabixFile, either on the server (default) or on the user's computer
all.col	TRUE to return all available information, FALSE to return a more compact result (the most informative columns, see Value)
file.export	Optional, name of the CSV file in which export the results (default is NULL)
n.cores	number of cores to use when scaning the TabixFile, can be efficient for large request (default is 1)

Value

The variants list with the assigned rfPred scores, as well as the scores used to build rfPred meta-score: SIFT, phyloP, MutationTaster, LRT (transformed) and Polyphen2 (corresponding to Polyphen2_HVAR_score). The data frame returned contains these columns:

chromosome	chromosome number
position_hg19	physical position on the chromosome as to hg19 (1-based coordinate)
reference	reference nucleotide allele (as on the + strand)
alteration	alternative nucleotide allele (as on the + strand)
proteine	Uniprot accession number
aaref	reference amino acid
aaalt	alternative amino acid
aapos	amino acid position as to the protein
rfPred_score	rfPred score betwen 0 and 1 (higher it is, higher is the probability of pathogenicity)
SIFT_score	SIFT score between 0 and 1 (higher it is, higher is the probability of pathogenicity contrary to the original SIFT score) = 1-original SIFT score
Polyphen2_score	Polyphen2 (HVAR one) score between 0 and 1, used to calculate rfPred (higher it is, higher is the probability of pathogenicity)
MutationTaster_score	MutationTaster score between 0 and 1 (higher it is, higher is the probability of pathogenicity)
PhyloP_score	PhyloP score between 0 and 1 (higher it is, higher is the probability of pathogenicity): PhyloP_score=1-0.5x10^phyloP if phyloP>0 or PhyloP_score=0.5x10^-phyloP if phyloP<0
LRT_score	LRT score between 0 and 1 (higher it is, higher is the probability of pathogenicity): LRT_score=1-LRToriginalx0.5 if LRT_Omega<1 or LRT_score=LRToriginalx0.5 if LRT_Omega>=1

The following columns are also returned if all.col is TRUE:

Uniprot_id	Uniprot ID number
genename	gene name
position_hg18	physical position on the chromosome as to hg18 (1-based coordinate)
Polyphen2_HDIV_score	Polyphen2 score based on HumDiv, i.e. hdiv_prob. The score ranges from 0 to 1: the corresponding prediction is "probably damaging" if it is in [0.957,1]; "possibly damaging" if it is in [0.453,0.956]; "benign" if it is in [0,0.452]. Score cut-off for binary classification is 0.5, i.e. the prediction is "neutral" if the score is lower than 0.5 and "deleterious" if the score is higher than 0.5. Multiple entries separated by ";"
Polyphen2_HDIV_pred	Polyphen2 prediction based on HumDiv: D (probably damaging), P (possibly damaging) and B (benign). Multiple entries separated by ";"
Polyphen2_HVAR_score	Polyphen2 score based on HumVar, i.e. hvar_prob. The score ranges from 0 to 1, and the corresponding prediction is "probably damaging" if it is in [0.909,1]; "possibly damaging" if it is in [0.447,0.908]; "benign" if it is in [0,0.446]. Score cut-off for binary classification is 0.5, i.e. the prediction is "neutral" if the score is lower than 0.5 and "deleterious" if the score is higher than 0.5. Multiple entries separated by ";"
Polyphen2_HVAR_pred	Polyphen2 prediction based on HumVar: D (probably damaging), P (possibly damaging) and B (benign). Multiple entries separated by ";"
MutationTaster_pred	MutationTaster prediction: A (disease_causing_automatic), D (disease_causing), N (polymorphism) or P (polymorphism_automatic)
phyloP	original phyloP score
LRT_Omega	estimated nonsynonymous-to-synonymous-rate ratio
LRT_pred	LRT prediction, D(eleterious), N(eutral) or U(nknown)

Author(s)

Fabienne Jabot-Hanin, Hugo Varet and Jean-Philippe Jais

References

Jabot-Hanin F, Varet H, Tores F and Jais J-P. 2013. rfPred: a new meta-score for functional prediction of missense variants in human exome (submitted).

Examples

# from a data.frame without uniprot protein identifier
data(variant_list_Y)
res=rfPred_scores(variant_list = variant_list_Y[,1:4],
                  data = system.file("extdata", "chrY_rfPred.txtz", package="rfPred",mustWork=TRUE),
                  index = system.file("extdata", "chrY_rfPred.txtz.tbi", package="rfPred",mustWork=TRUE))
# from a data.frame with uniprot protein identifier
res2=rfPred_scores(variant_list = variant_list_Y,
                   data = system.file("extdata", "chrY_rfPred.txtz", package="rfPred",mustWork=TRUE),
                   index = system.file("extdata", "chrY_rfPred.txtz.tbi", package="rfPred",mustWork=TRUE))
# from a VCF file
res3=rfPred_scores(variant_list = system.file("extdata", "example.vcf", package="rfPred",mustWork=TRUE),
                   data = system.file("extdata", "chrY_rfPred.txtz", package="rfPred",mustWork=TRUE),
                   index = system.file("extdata", "chrY_rfPred.txtz.tbi", package="rfPred",mustWork=TRUE))
# from a GRanges object
data(example_GRanges)
res4=rfPred_scores(variant_list = example_GRanges,
                   data = system.file("extdata", "chrY_rfPred.txtz", package="rfPred",mustWork=TRUE),
                   index = system.file("extdata", "chrY_rfPred.txtz.tbi", package="rfPred",mustWork=TRUE))

---

Motor of rfPred_scores

Description

Motor of rfPred_scores

Usage

rfPred_scores_motor(variant_list, data, index, all.col, file.export, n.cores)

Arguments

variant_list	Variants list in a data.frame containing 4 or 5 columns: chromosome number, hg19 genomic position on the chromosome, reference nucleotid, variant nucleotid and uniprot protein identifier (optional)
data	Path to the compressed TabixFile, either on the server (default) or on the user's computer
index	Path to the index of the TabixFile, either on the server (default) or on the user's computer
all.col	TRUE to return all available information, FALSE to return a more compact result (the most informative columns, see Value)
file.export	Optional, name of the CSV file in which export the results (default is NULL)
n.cores	number of cores to use when scaning the TabixFile, can be efficient for large request (default is 1)

Value

see the rfPred_scores function

Note

This function is called by the rfPred_scores S4 method

---

Toy example of data.frame

Description

Toy example of data.frame

Format

A data frame with 5 observations on the following 5 variables:

chr

Chromosome number (only Y in this example)

pos

Position on the chromosome (numeric)

ref

Referent nucleotid (A, C, G or T)

alt

Alteration nucleotid (A, C, G or T)

uniprot

Uniprot protein identifier (factor)

---