NEWS.md
plot_signalp() has been deprecated and removed.
get_signalp5() now works with https://services.healthtech.dtu.dk/service.php?SignalP-5.0 site since the old link http://www.cbs.dtu.dk/services/SignalP/ does not function any longer.
get_signalp() now works with https://services.healthtech.dtu.dk/service.php?SignalP-4.1 since the old link http://www.cbs.dtu.dk/services/SignalP-4.1/ does not function any longer.
get_signalp() now now runs one job at a time.
get_signalp() splitter argument default value has been changed to 1000.
get_signalp() sleep argument has been removed.
get_targetp() now works with https://services.healthtech.dtu.dk/service.php?TargetP-1.1 since the old link http://www.cbs.dtu.dk/services/TargetP-1.1/ does not function any longer.
get_targetp() now now runs one job at a time.
get_targetp() splitter argument default value has been changed to 1000.
get_targetp() sleep argument has been removed. .
get_cdd() which queries the the Conserved Domain Database (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi)get_signalp5() which queries SignalP5 web server (http://www.cbs.dtu.dk/services/SignalP)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. Data.frame input is accepted as well.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. Data.frame input is accepted as well.plot_prot() domain argument can now be "cdd", "hmm" or "none". Default is "cdd". This argument determines if get_hmm() or get_cdd() are used for domain annotation. Data.frame input is accepted as well.get_* and scan_* functions, as well as maab() now work with AAStringSet class objects. #5get_signalp() output contains an additional column sp.length - integer, length of the predicted signal peptide. This column is a copy of Ymax.pos. Potentially BREAKING CHANGEget_phobius() output contains an additional column sp.length - integer, length of the predicted signal peptide. Potentially BREAKING CHANGE
get_phobius() the name of the output column “Name” has been changed to “id”. BREAKING CHANGE
plot_prot() default value for gpi argument has been changed to “netgpi”.get_netGPI() now now runs one job at a time.
get_netGPI() splitter argument default value has been changed to 2500.
get_targetp() has been fixed to work with org_type = "non_plant".
get_espritz() has been restored since Espritz server is again available at http://old.protein.bio.unipd.it/espritz/.
plot_prot() has new arguments gpi_size - controls the size of the gpi symbol and gpi_shape- controls the shape of the gpi symbol.
plot_prot() has new argument hyp_scan - which is a logical and determines if scan_ag() (when ag = TRUE) should scan for arabinogalactan motifs containing only predicted hydroxyprolines. This argument changes the previous default behavior of plot_prot() which was equivalent to hyp_scan = FALSE.
get_phobius() and get_big_pi() now use https web server addresses. #6
get_espritz() has been removed since Espritz server is no longer available for obtaining predictions. BREAKING CHANGE.plot_prot() arguments nsp, domain, tm, gpi and disorder can also be user supplied data frames obtained by calling get_signalp(), get_hmm(), get_phobius(), get_big_pi(), get_pred_gpi(), get_netGPI() and get_espritz() on the same sequences supplied to plot_prot().get_big_pi() output column is.bigpi has been renamed to is.gpi to bring the output in line with other gpi predicting functions. BREAKING CHANGE.get_big_pi() output column order has been changed to: id, is.gpi, Quality, omega_site and PValue. BREAKING CHANGE.New get_netGPI() queries NetGPI web server (https://services.healthtech.dtu.dk/service.php?NetGPI-1.0) for predictions of GPI-anchored proteins.
maab() argument get_gpi has an additional option "netgpi". When set, the function will query NetGPI web server to resolve ambiguities in maab classes depending on GPI-anchoring predictions.
plot_prot() argument gpi has an additional option "netgpi". When set, the function will query NetGPI web server for prediction of omega sites.
maab() now correctly performs when only a single protein sequence is provided as an argument.
get_hmm() receives additional numeric arguments: ievalue and bitscore. These arguments are used to filter sequences with lower or equal ievalue and higher or equal bitscore in the output. This is useful when used from plot_prot() to avoid plotting weakly identified domains.
get_big_pi() and get_pred_gpi() now return NA in respective is.bigpi and is.gpi columns when the servers are unable to make a prediction due to non-amino acid letters or length of the sequence.
maab() now correctly does not resolve class ambiguities when the logical vector provided as gpi argument contains NA values. Previously it returned NA as maab_class. This is useful when maab() is called with get_gpi = 'predgpi' or get_gpi = 'bigpi' arguments, and the corresponding servers are unable to make a prediction for a sequence due to non-amino acid letters or length of the sequence.
predict_hyp() internal model is updated to 2nd version (‘V2’). Predictions are around 25% faster compared to the first version. The performance in terms of accuracy is similar based on the test set used. ‘V2’ was created using a more streamlined manner and is the default model. The old model (‘V1’) is still available using the version argument to predict_hyp().predict_hyp() now checks if all provided ids are unique. Previously non unique ids caused an error at the end of computation.
predict_hyp() sequence output has changed for sequences containing non amino acid letters. Previously NA was returned for such sequences. At present all “P”" for which the probability is higher then the defined threshold (tprob argument) are changed to “O”" and all others are left as “P”.
maab() now correctly outputs when there are no MAAB classes found and get_gpi argument is set to "predgpi" or "bigpi". Previously this caused an error due to stringr::write.fasta() attempting to write a file with no sequences.
pfam2go() now takes Pfam > GO mappings from ftp://ftp.geneontology.org/pub/go/external2go/pfam2go instead of http://geneontology.org/external2go/pfam2go.
get_phobius(), get_big_pi(), get_pred_gpi(), maab() and plot_prot(), gain an additional argument progress. progress is a logical value determining whether to show the progress bar, (default FALSE).
get_targetp(), get_signalp() and get_hmm() argument progress is now set to FALSE at default.
get_hmm() argument verbose is now set to FALSE at default.
added pkgdown site for ragp at: https://missuse.github.io/ragp/.
get_targetp() and get_signalp() gain additional arguments progress and attempts. progress is a logical value determining whether to show the progress bar, (default TRUE). attempts is an integer value determining the number of repeated attempts if server unresponsive (default 2). These functions now return finished queues if server becomes unresponsive.get_big_pi(), get_espritz(),get_hmm(), get_phobius(), get_pred_gpi(), get_signalp(), get_targetp(), maab(), predict_hyp(), scan_ag() and scan_nglc() now use the S3 object system which will make further extensions of accepted inputs straightforward.Completely rewrote the code for scan_ag() which is now simplified and easier to read. The function is now about 20% slower.
fixed a bug in scan_ag() when argument exclude_ext = "all" which resulted in detection of unwanted amino acids in certain sequence arrangements. The bug occurred only if AG glycomoduls were detected, it did not introduce any. If you used this function with exclude_ext = "all" it is advisable to rerun these analyses again.
removed deprecated functions: get_signap_file(), get_targetp_file(), get_phobius_file().
New get_espritz() queries ESpritz web server for predictions on protein disordered regions.
plot_prot() gains an additional argument disorder, logical indicating should the predicted disordered regions be plotted. Defaults to FALSE.
plot_prot() introduced in 0.1.0.0003 which prevented GPIs to be plotted when gpi = "bigpi".get_pred_gpi() queries PredGPI web server for predictions on GPI presence and omega site location.maab() argument get_gpi has been changed and now accepts strings as input: get_gpi = c("bigpi", "predgpi", "none") which indicate whether to query Big Pi or PredGPI server or not to resolve class ambiguities.
plot_prot() argument gpi has been changed and now accepts strings as input gpi = c("bigpi", "predgpi", "none") which indicate whether to query Big Pi or PredGPI server or not to plot GPI positions.
fixed a bug in plot_prot() which caused extra-cellular regions to start at 0 instead of 1.
fixed a bug in get_big_pi() when shorter than 55 amino acid sequences were provided.
When using plot_prot() with argument dom_sort = "ievalue" the domains with the lowest independent e-value will now be correctly plotted on top.
The ratio of x and y axes in plot_prot() output is now calculated based on sequence length and should provide more consistent diagrams.
get_hmm() can now handle protein sequences of arbitrary length by splitting them into several shorter overlapping sequences and querying hmmscan.
New plot_prot() returns a ggplot2 diagram of protein structure based on hmmscan domain annotation and several types of predictions.
New get_signalp(), get_targetp() and get_phobius() replace deprecated get_signalp_file(), get_targetp_file() and get_phobius_file(). These functions accept R objects as well as FASTA files as input (#5).
New plot_signalp() takes one single letter protein sequence and returns a detailed SignalP prediction along with a plot.
New scan_nglc() detects motifs for N-glycosylation on asparagine residues.
maab() gain an additional logical argument get_gpi (default FALSE); if TRUE get_big_pi will be called on all sequences belonging to one of the HRGP classes thus resolving class ambiguities that depend on GPI knowledge.
Added a NEWS.md file to track changes to the package.
Significantly improved the speed of get_big_pi(). Removed the verbose argument. Added a progress bar (#3).
get_hmm() gains new arguments: timeout - time in seconds to wait for the server response (default is 10s). attempts - number of attempts if server unresponsive (default is 2 times) (#3).
get_signalp() and get_targetp() now perform 10 parallel queries instead of arbitrary many. This results in higher stability when running on many sequences, and a slight reduction in speed. These functions now have a progress bar.
get_phobius() gains a progress bar.
Changed internal behavior when trunc argument is specified for get_signalp(). Now the truncation is performed prior to sending the files to the server resulting in higher efficiency.
scan_ag() gains logical tidy argument. The specific output is available when simplify = FALSE and tidy = TRUE in function call (#1).