About
Click here
to learn more about this blog.
Packages
Click on a link
for updates to that package only.
Subscribe
Subscribe to the
Bioconductor Changlog.
Thu, 02 Jul 2009
Biostrings
Use closest integer mismatch in trimLRPatterns when mismatch parameter is in (0,1).
Biostrings
Bumped version to 2.13.22 in order to release the new code for fast matching of the head/tail of a PDict object. In some situations where the head/tail is very narrow (<= 8b in total), the overall speed up for matchPDict() can be 10x or more.
Biostrings
First working version of the preprocessing and fast matching of the head and tail of a PDict object. Still needs some testing.
Biostrings
Added the N50() function from Nicolas Delhomme. Bumped version to 2.13.21
Biostrings
Third (and hopefully last) attempt to refactor the match reporting
facilities used by matchPDict() at the C level. This one looks good.
The design of the new code should make it easy to support storage of
matches that have a width different from the width of the pattern
they derive from (this will happen when matchPDict() will support
indels). Bumped version to 2.13.20
Temporarily reverting back to r40157.
Second attempt to refactor the match reporting facilities used by
matchPDict() at the C level. Code is still not ready (will revert
back in the next commit).
Resync with latest change to the "Auto-Extending buffer" code in IRanges. Bumped version to 2.13.19
Biostrings
Added more functionality to support position weight matrix activities.
Biostrings
Added headers for R math and random number generation.
Temporarily reverting back to r40103.
Biostrings
Temporarily reverting back to r40103.
Refactored the match reporting facilities used by matchPDict() at the
C level: code is now simpler and cleaner (no more global variable).
Bumped version to 2.13.18
Improving on dinucleotideFrequencyTest output.
Adding Monte Carlo simulations to dinucleotideFrequencyTest.
Removing unneeded variables for pairwise alignment C code.
Biostrings
The first draft of a matching method based upon weighted clustered positions (WCP), a conceptual generalization of matching based on position weight matrices. This new scheme uses three new categories of classes: XKeySortedData, XKeySortedDataList, and WCP. This check in also made some improvements to maxScore/maxWeights (PWM) and XStringSet infrastructure.
Fixed a bug in pairwiseAlignment when scoreOnly = TRUE and gapOpening = 0.
Biostrings
Adjusted hard-coded limit in order to support 'max.nedit' values up to 100. Bumped version to 2.13.16
Fixed a comment in XStringViews_match_PWM
Biostrings
Rename internal C function compute_score to compute_pwm_score to avoid confusion with future score computing functions.
Cleaned up comments in matchPWM method for XStringViews objects.
Biostrings
Added matchPWM and countPWM methods for XStringViews and MaskedDNAString objects.
Biostrings
Speed up "countIndex" method for MIndex objects (by using new
"elementLengths" method for list objects).
Bumped version to 2.13.14
Biostrings
Added an is.unsorted method for XStringSet objects and removed XStringSet_in_set since the cost of its maintenance outweighed the speedup it had over using match to produce %in% output.
Make NCBI download of BLOSUM matrix condition on being interactive to avoid unimportant R CMD check errors.
Small adjustments to the recent changes/additions:
- MIndex now contains IRangesList (more accurate than RangesList).
- Removed XStringSet_not_duplicated C code (not needed, having
the XStringSet_duplicated C code is enough). The same would
probably apply to the XStringSet_in_set/XStringSet_match C code
(seems redundant).
- Fixed segfault in "rank" method for XStringSet objects when
the object has no elements.
Added doc for fastq.geometry().
Fixed a bug in duplicated method for XStringSet objects when length = 1.
Biostrings
Fixed a bug in duplicated method for XStringSet objects when length = 1.
Fixed vignette usepackage issue on Windows.
Switched to using binary search in %in% and match methods for XStringSet objects.
Allow for single numeric nomatch arg, which will be converted to integer, to match method for XStringSet objects.
Biostrings
Allow for single numeric nomatch arg, which will be converted to integer, to match method for XStringSet objects.
Created no-copy match methods for XStringSet objects.
Adding some comments to code.
Fixed the order method for XStringSet to coincide with order method in R by performing a primary qsort on value and secondary qsort on index.
Improved upon %in% method for XStringSet objects.
Fixed match function for XStringSet objects.
Minor code formatting change.
Added "%in%" and match methods for XStringSet objects.
Biostrings
A long due change (better late than never): renamed the 'freq'
argument 'as.prob' in the following functions:
alphabetFrequency
oligonucleotideFrequency
dinucleotideFrequency
trinucleotideFrequency
oligonucleotideTransitions
nucleotideFrequencyAt
consensusMatrix
For backward compatibility, 'freq' is still accepted but a warning
is issued if it's used (this will be removed during Biostrings 2.15
devel cycle).
Bumped version to 2.13.10
[merge -r 39724:39725 from release version]
Put call to pairwiseAlignment() in an "interactive() block" in man page
for matchPattern(). This is to avoid the 'R CMD check' failure we get on
our Windows build machines because the 733.5 Mb of memory needed by this
call cannot be allocated.
Removed "length" method for XString objects (no need to override "length" method for XSequence objects).
Biostrings
Major refactor to the IRanges package. The class system became unwieldy and we decided to make the Sequence virtual class the parent of nearly all classes in this package, much like the vector virtual class is the parent of nearly all classes in the S3 world.
Biostrings
Modified replaceLetterAt() example.
Some improvements to replaceLetterAt():
- added support for logical or Rle 'at' argument in "replaceLetterAt"
method for DNAString objects;
- added a "replaceLetterAt" method for DNAStringSet objects (with the
current restrictions that 'x' and 'at' must be rectangular i.e. 'x'
must have a constant width and 'at' must be a logical matrix with
the same dimensions as 'x').
Bumped version to 2.13.8
Fixed coverage method for PairwiseAlignedFixedSubjectSummary objects.
Biostrings
Added "unlist" method for XStringSet objects.
- Added "subseq" and "subseq<-" methods for character vectors and
XStringSet objects.
- Bumped version to 2.13.7
Updated dependents of IRanges to reflect changes in List/ListLike classes.
Biostrings
Updated dependents of IRanges to reflect changes in List/ListLike classes.
Biostrings
Added the compact() generic. Bumped version to 2.13.5
Biostrings
Added a TypedListV2 (TypedList version 2) class to replace TypedList virtual class. It separate implementation from concept and includes the usage of SimpleTypedList and CompressedTypedList virtual classes.
Biostrings
Resync with latest changes in the Partitioning class in the IRanges
package. Bumped version to 2.13.3
Biostrings
Added an example showing an alternative algorithm for mapping probe set IDs between chips (has even more restrictions than the first algorithm).
Improved vwhichPDict() speed by using vmembers() internally for
adding the ids of the duplicated patterns to the result.
Added a full example showing how to use it to map probe set IDs
between chips.
Bumped version to 2.13.2.
Biostrings
Improved vwhichPDict() speed by using vmembers() internally for
adding the ids of the duplicated patterns to the result.
Added a full example showing how to use it to map probe set IDs
between chips.
Bumped version to 2.13.2.
Biostrings
Removed local Dups class and use Dups class from the IRanges package.
Added vwhichPDict() (not yet optimized). Also whichPDict() now returns the pattern indices in ascending order.
Biostrings
matchPWM(): When passed a Position Weight Matrix of doubles, matchPWM()
doesn't coerce it to an integer matrix anymore. The same apply to its
'min.score' argument. PWMscoreStartingAt() now always returns a vector
of doubles, whatever the storage mode of the PWM was (this could be
Bumped version to 2.13.1
Biostrings
added doc alias for Expanded_TB_PDict
Biostrings
added doc alias for Expanded_TB_PDict
XStringSetList now contains IPartitioning
added stub for man/XStringSetList-class.Rd
Added some basic operations on XStringSetList objects (not intended to be a full featured class for now).
Added the XStringSetList (virtual), BStringSetList, DNAStringSetList, RNAStringSetList, AAStringSetList and Expanded_TB_PDict classes. No methods yet.
Biostrings
Bumped pkg versions for BioC 2.5 devel line.
Bumped pkg versions for BioC 2.4 release.
Added "width" method for MIndex objects.
Renamed ByName_MIndex class -> SparseMIndex and completely turned off
any code related to this class (the SparseMIndex class is not used
anymore, but might be needed again in the future). As a side effect,
'R CMD check' is now totally happy.
Fixed bug in consensusMatrix() when some of the values passed thru the 'shift' arg were <= - 'width'.
small improvement to man/nucleotideFrequency.Rd
Biostrings
Added nucleotideFrequencyAt() (see ?nucleotideFrequencyAt). Bumped version to 2.11.56
Biostrings
Added hasLetterAt() (see ?hasLetterAt).
Biostrings
Added "reverse" method for character objects and deprecated strrev() in its favor.
Moved oligonucleotideFrequency() and related functions to its own man page.
oligonucleotideFrequency(): implementation is now entirely in C (except argument checking). Bumped version to 2.11.55
Biostrings
oligonucleotideFrequency(): some code optimizations.
oligonucleotideFrequency(): the "shape" of the answer is now controlled
thru the 'as.array' and 'simplify.as' arguments.
oligonucleotideFrequency():
- Replaced 'as.array' arg by 'as' arg.
- Return a matrix by default on an XStringSet object.
- Moved top-level loop in method for XStringSet objects to the C level
(except when 'as' is not "atomic" and 'collapse' is FALSE).
Biostrings
Added 'collapse' and 'weight' arguments to vcountPDict().
Bumped version to 2.11.54
Biostrings
Change TypedList class to inherit from ListLike class.
Biostrings
Some reformatting of man/XStringViews-class.Rd
Use the new Views() interface.
Bumped version to 2.11.52
Biostrings
cleaned Depends field
Biostrings
Got rid of load-time warnings:
Warning in namespaceImportFrom(self, asNamespace(ns)) :
replacing previous import: head
Warning in namespaceImportFrom(self, asNamespace(ns)) :
replacing previous import: tail
Bumped version to 2.11.51
"consensusMatrix" method for XStringSet objects: added some details
about the 'width=NULL' mode
Make use of the new coverage() interface. Bumped version to 2.11.50