AAStringSet support

DESCRIPTION

@@ -17,14 +17,17 @@ Description: Identification of plant hydroxyproline rich glycoprotein sequences
 Depends: R (>= 3.3.0)
 License: file LICENSE
 LazyData: true
-Imports: xgboost (>= 1.1.1.1),
+Imports: 
+    xgboost (>= 1.1.1.1),
     stringr (>= 1.2.0),
     httr (>= 1.3.1),
     xml2 (>= 1.1.1),
     seqinr(>= 3.3),
     utils (>= 3.3),
     graphics (>= 3.3),
-    ggplot2 (>= 2.2.1)
+    ggplot2 (>= 2.2.1),
+    Biostrings (>= 2.6),
+	BiocGenerics (>= 0.3)
 RoxygenNote: 7.1.1
 Suggests:
     protr,
@@ -37,3 +40,6 @@ URL: https://github.com/missuse/ragp
 BugReports: https://github.com/missuse/ragp/issues
 Encoding: UTF-8
 VignetteBuilder: knitr
+Remotes:
+    bioc::release/Biostrings
+	bioc::release/BiocGenerics

NAMESPACE

@@ -1,57 +1,71 @@
 # Generated by roxygen2: do not edit by hand
 
+S3method(get_big_pi,AAStringSet)
 S3method(get_big_pi,character)
 S3method(get_big_pi,data.frame)
 S3method(get_big_pi,default)
 S3method(get_big_pi,list)
+S3method(get_espritz,AAStringSet)
 S3method(get_espritz,character)
 S3method(get_espritz,data.frame)
 S3method(get_espritz,default)
 S3method(get_espritz,list)
+S3method(get_hmm,AAStringSet)
 S3method(get_hmm,character)
 S3method(get_hmm,data.frame)
 S3method(get_hmm,default)
 S3method(get_hmm,list)
+S3method(get_netGPI,AAStringSet)
 S3method(get_netGPI,character)
 S3method(get_netGPI,data.frame)
 S3method(get_netGPI,default)
 S3method(get_netGPI,list)
+S3method(get_phobius,AAStringSet)
 S3method(get_phobius,character)
 S3method(get_phobius,data.frame)
 S3method(get_phobius,default)
 S3method(get_phobius,list)
+S3method(get_pred_gpi,AAStringSet)
 S3method(get_pred_gpi,character)
 S3method(get_pred_gpi,data.frame)
 S3method(get_pred_gpi,default)
 S3method(get_pred_gpi,list)
+S3method(get_signalp,AAStringSet)
 S3method(get_signalp,character)
 S3method(get_signalp,data.frame)
 S3method(get_signalp,default)
 S3method(get_signalp,list)
+S3method(get_signalp5,AAStringSet)
 S3method(get_signalp5,character)
 S3method(get_signalp5,data.frame)
 S3method(get_signalp5,default)
 S3method(get_signalp5,list)
+S3method(get_targetp,AAStringSet)
 S3method(get_targetp,character)
 S3method(get_targetp,data.frame)
 S3method(get_targetp,default)
 S3method(get_targetp,list)
+S3method(get_tmhmm,AAStringSet)
 S3method(get_tmhmm,character)
 S3method(get_tmhmm,data.frame)
 S3method(get_tmhmm,default)
 S3method(get_tmhmm,list)
+S3method(maab,AAStringSet)
 S3method(maab,character)
 S3method(maab,data.frame)
 S3method(maab,default)
 S3method(maab,list)
+S3method(predict_hyp,AAStringSet)
 S3method(predict_hyp,character)
 S3method(predict_hyp,data.frame)
 S3method(predict_hyp,default)
 S3method(predict_hyp,list)
+S3method(scan_ag,AAStringSet)
 S3method(scan_ag,character)
 S3method(scan_ag,data.frame)
 S3method(scan_ag,default)
 S3method(scan_ag,list)
+S3method(scan_nglc,AAStringSet)
 S3method(scan_nglc,character)
 S3method(scan_nglc,data.frame)
 S3method(scan_nglc,default)

NEWS.md

@@ -6,8 +6,9 @@ New Features
 
 * added new function `get_signalp5()` which queries SignalP5 web server (http://www.cbs.dtu.dk/services/SignalP)  
 * added new function `get_tmhmm()` which queries TMHMM  v. 2.0 web server (http://www.cbs.dtu.dk/services/TMHMM/)  
-* `plot_prot()` `nsp` argument can now be `"signalp"`, `"signalp5"` or `"none"`. Default is `"signalp5"`. This argument determines if `get_signalp()` or `get_signalp5()` are used for N-sp prediction. **BREAKING CHANGE**
-* `plot_prot()` `tm` argument can now be `"phobius"`, `"tmhmm"` or `"none"`. Default is `"phobius"`. This argument determines if `get_phobius()` or `get_tmhmm()` are used for TM prediction.  **BREAKING CHANGE**
+* `plot_prot()` `nsp` argument can now be `"signalp"`, `"signalp5"` or `"none"`. Default is `"signalp5"`. This argument determines if `get_signalp()` or `get_signalp5()` are used for N-sp prediction. 
+* `plot_prot()` `tm` argument can now be `"phobius"`, `"tmhmm"` or `"none"`. Default is `"phobius"`. This argument determines if `get_phobius()` or `get_tmhmm()` are used for TM prediction.  
+* all `get_*` and `scan_*` functions, as well as `maab()` now work with `AAStringSet` class objects. #5
 
 Bug Fixes and Improvements
 --------------------------

R/get_big_pi.R

@@ -2,8 +2,8 @@
 #'
 #' big-PI Plant Predictor is a web server utilizing a scoring algorithm for prediction of GPI modification sites in plants.
 #'
-#' @aliases get_big_pi get_big_pi.default get_big_pi.character get_big_pi.data.frame get_big_pi.list
-#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Should be left blank if vectors are provided to sequence and id arguments.
+#' @aliases get_big_pi get_big_pi.default get_big_pi.character get_big_pi.data.frame get_big_pi.list get_big_pi.AAStringSet
+#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Alternatively an `AAStringSet` object. Should be left blank if vectors are provided to sequence and id arguments.
 #' @param sequence An appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param id An appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param simplify A boolean indicating the type of returned object, defaults to TRUE.
@@ -502,3 +502,25 @@ get_big_pi.default <- function(data = NULL,
   }
   return(res_out)
 }
+
+#' @rdname get_big_pi
+#' @method get_big_pi AAStringSet
+#' @export
+
+
+get_big_pi.AAStringSet <-  function(data,
+                                    ...){
+  sequence <- as.character(data)
+  id <- names(sequence)
+  sequence <- unname(sequence)
+  sequence <- toupper(sequence)
+  sequence <- sub("\\*$",
+                  "",
+                  sequence)
+
+  res <- get_big_pi.default(sequence = sequence,
+                            id = id,
+                            ...)
+  return(res)
+}
+

R/get_espritz.R

@@ -2,8 +2,8 @@
 #'
 #' Espritz web server predicts disordered regions from primary sequence. It utilizes Bi-directional Recursive Neural Networks and can process proteins on a genomic scale with little effort and state-of-the-art accuracy.
 #'
-#' @aliases get_espritz get_espritz.default get_espritz.character get_espritz.data.frame get_espritz.list
-#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Should be left blank if vectors are provided to sequence and id arguments.
+#' @aliases get_espritz get_espritz.default get_espritz.character get_espritz.data.frame get_espritz.list get_espritz.AAStringSet
+#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Alternatively an `AAStringSet` object. Should be left blank if vectors are provided to sequence and id arguments.
 #' @param sequence A vector of strings representing protein amino acid sequences, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param id A vector of strings representing protein identifiers, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param model One of c('X-Ray', 'Disprot', 'NMR'), default is 'X-Ray'. Determines the model to be used for prediction. See details.
@@ -406,3 +406,23 @@ get_espritz.default <- function(data = NULL,
                  res)
   return(res)
 }
+
+#' @rdname get_espritz 
+#' @method get_espritz AAStringSet
+#' @export
+
+get_espritz.AAStringSet <-  function(data,
+                                     ...){
+  sequence <- as.character(data)
+  id <- names(sequence)
+  sequence <- unname(sequence)
+  sequence <- toupper(sequence)
+  sequence <- sub("\\*$",
+                  "",
+                  sequence)
+
+  res <- get_espritz.default(sequence = sequence,
+                             id = id,
+                             ...)
+  return(res)
+}

R/get_hmm.R

@@ -3,8 +3,8 @@
 #' hmmer web server offers biosequence analysis using profile hidden Markov Models. This function allows searching
 #' of a protein sequence vs a profile-HMM database (Pfam-A).
 #'
-#' @aliases get_hmm get_hmm.default get_hmm.character get_hmm.data.frame get_hmm.list
-#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Should be left blank if vectors are provided to sequence and id arguments.
+#' @aliases get_hmm get_hmm.default get_hmm.character get_hmm.data.frame get_hmm.list get_hmm.AAStringSet
+#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Alternatively an `AAStringSet` object. Should be left blank if vectors are provided to sequence and id arguments.
 #' @param sequence A vector of strings representing protein amino acid sequences, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param id A vector of strings representing protein identifiers, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param verbose Boolean, whether to print out the output for each sequence, defaults to FALSE.
@@ -556,3 +556,23 @@ get_hmm.default <- function(data = NULL,
   rownames(pfam) <- NULL
   return(pfam)
 }
+
+#' @rdname get_hmm 
+#' @method get_hmm AAStringSet
+#' @export
+
+get_hmm.AAStringSet <-  function(data,
+                                 ...){
+  sequence <- as.character(data)
+  id <- names(sequence)
+  sequence <- unname(sequence)
+  sequence <- toupper(sequence)
+  sequence <- sub("\\*$",
+                  "",
+                  sequence)
+
+  res <- get_hmm.default(sequence = sequence,
+                         id = id,
+                         ...)
+  return(res)
+}

R/get_netGPI.R

@@ -2,8 +2,8 @@
 #'
 #' NetGPI server offers GPI Anchor predictions
 #'
-#' @aliases get_netGPI get_netGPI.default get_netGPI.character get_netGPI.data.frame get_netGPI.list
-#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Should be left blank if vectors are provided to sequence and id arguments.
+#' @aliases get_netGPI get_netGPI.default get_netGPI.character get_netGPI.data.frame get_netGPI.list get_netGPI.AAStringSet
+#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Alternatively an `AAStringSet` object. Should be left blank if vectors are provided to sequence and id arguments.
 #' @param sequence A vector of strings representing protein amino acid sequences, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param id A vector of strings representing protein identifiers, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param splitter An integer indicating the number of sequences to be in each .fasta file that is to be sent to the server. Defaults to 2500. Change only in case of a server side error. Accepted values are in range of 1 to 5000.
@@ -535,3 +535,23 @@ get_netGPI.default <- function(data = NULL,
   res <- get_netGPI.character(data = file_name, ...)
   return(res)
 }
+
+#' @rdname get_netGPI 
+#' @method get_netGPI AAStringSet
+#' @export
+
+get_netGPI.AAStringSet <-  function(data,
+                                    ...){
+  sequence <- as.character(data)
+  id <- names(sequence)
+  sequence <- unname(sequence)
+  sequence <- toupper(sequence)
+  sequence <- sub("\\*$",
+                  "",
+                  sequence)
+
+  res <- get_netGPI.default(sequence = sequence,
+                            id = id,
+                            ...)
+  return(res)
+}

R/get_phobius.R

@@ -2,8 +2,8 @@
 #'
 #' Phobius web server is a combined transmembrane topology and signal peptide (N-sp) predictor. Currently only "normal prediction" of signal peptides is supported by the function.
 #'
-#' @aliases get_phobius get_phobius.default get_phobius.character get_phobius.data.frame get_phobius.list
-#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Should be left blank if vectors are provided to sequence and id arguments.
+#' @aliases get_phobius get_phobius.default get_phobius.character get_phobius.data.frame get_phobius.list get_phobius.AAStringSet
+#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Alternatively an `AAStringSet` object. Should be left blank if vectors are provided to sequence and id arguments.
 #' @param sequence A vector of strings representing protein amino acid sequences, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param id A vector of strings representing protein identifiers, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param progress Boolean, whether to show the progress bar, at default set to FALSE.
@@ -305,3 +305,23 @@ get_phobius.default <- function(data = NULL,
                                ...)
   return(res)
 }
+
+#' @rdname get_phobius
+#' @method get_phobius AAStringSet
+#' @export
+
+get_phobius.AAStringSet <-  function(data,
+                                     ...){
+  sequence <- as.character(data)
+  id <- names(sequence)
+  sequence <- unname(sequence)
+  sequence <- toupper(sequence)
+  sequence <- sub("\\*$",
+                  "",
+                  sequence)
+
+  res <- get_phobius.default(sequence = sequence,
+                             id = id,
+                             ...)
+  return(res)
+}

R/get_pred_gpi.R

@@ -2,8 +2,8 @@
 #'
 #' PredGPI web server is a predictor of GPI modification sites.
 #'
-#' @aliases get_pred_gpi get_pred_gpi.default get_pred_gpi.character get_pred_gpi.data.frame get_pred_gpi.list
-#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Should be left blank if vectors are provided to sequence and id arguments.
+#' @aliases get_pred_gpi get_pred_gpi.default get_pred_gpi.character get_pred_gpi.data.frame get_pred_gpi.list get_pred_gpi.AAStringSet
+#' @param data A data frame with protein amino acid sequences as strings in one column and corresponding id's in another. Alternatively a path to a .fasta file with protein sequences. Alternatively a list with elements of class "SeqFastaAA" resulting from \code{\link[seqinr]{read.fasta}} call. Alternatively an `AAStringSet` object. Should be left blank if vectors are provided to sequence and id arguments.
 #' @param sequence A vector of strings representing protein amino acid sequences, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param id A vector of strings representing protein identifiers, or the appropriate column name if a data.frame is supplied to data argument. If .fasta file path, or list with elements of class "SeqFastaAA" provided to data, this should be left blank.
 #' @param spec Numeric in the 0-1 range, indicating the threshold specificity.
@@ -348,4 +348,22 @@ get_pred_gpi.default <- function(data = NULL,
   return(res)
 }
 
+#' @rdname get_pred_gpi
+#' @method get_pred_gpi AAStringSet
+#' @export
 
+get_pred_gpi.AAStringSet <-  function(data,
+                                      ...){
+  sequence <- as.character(data)
+  id <- names(sequence)
+  sequence <- unname(sequence)
+  sequence <- toupper(sequence)
+  sequence <- sub("\\*$",
+                  "",
+                  sequence)
+
+  res <- get_pred_gpi.default(sequence = sequence,
+                              id = id,
+                              ...)
+  return(res)
+}