rfPred_scores {rfPred} | R Documentation |
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.
variant_list |
A variants list in a
|
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 |
|
file.export |
Optional, name of the CSV file in
which export the results (default is |
n.cores |
number of cores to use when scaning the TabixFile, can be efficient for large request (default is 1) |
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: |
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: |
MutationTaster_pred |
MutationTaster prediction:
|
phyloP |
original phyloP score |
LRT_Omega |
estimated nonsynonymous-to-synonymous-rate ratio |
LRT_pred |
LRT prediction, |
Fabienne Jabot-Hanin, Hugo Varet and Jean-Philippe Jais
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).
# 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))