New Features and History of Changes
Here are some of the more notable new features and bug fixes
since version 1.0. Bugs that would have yielded errors in substantive
results (e.g. a bug in a phylogenetic calculation) are marked
with  .
Beginning with version 1.1, a list
of new modules in the latest version can be seen by choosing "List
new modules" from the Help menu.
Version 1.11
Bug Fixes
- ()
In previous versions of Mesquite, a Taxa Association between
taxa blocks, for example one which records what haplotypes belong
to what species, could come to have incorrect linkings if taxa
were reordered in the file. This is not directly a substantive
bug, but could yield calculations with errors if the user does
not notice that taxa have become mis-associated. This bug has
been fixed
- Clustal Align now works under Windows
- Various other small bugs were fixed.
Version 1.1
New Features
- Character evolution
- Gene tree/species trees
- Alignment (see
the
Align Package manual)
- Manual alignment tools
in editor (Block Mover, Sequence Pusher)
- Automated pairwise alignment
tool in editor
- Submission of a selected region
in the matrix to be aligned by ClustalW and
then reincorporation of the newly aligned region into the
matrix
- Genetic
Codes
- Assignment of genetic
codes to individual characters
- Translation
of nucleotide data to protein data
- Molecular Data
- Distances
- Jukes-Cantor, Felsenstein 81,
Kimura 2-Parameter, Felsenstein 84 distances
- Great Circle distances in the
separately-released Cartographer
package
- Charts
- Values for Taxon Pairs now available
in Bar & Line Charts and Scattergrams, allowing one to
plot, for example, uncorrected distance versus corrected distance
for all taxon pairs within a matrix.
- Line charts
- Bar & Line charts can now
show cumulative results in various fashions
(e.g., cumulative, cumulative average, reverse cumulative
average)
- Interactions
with GenBank
- Import into matrix the top BLAST
matches to sequences selected in matrix
- Interaction with MrBayes
- live updating
of trees produced by ongoing MrBayes analysis
- exporter that merges
matrices and includes a MrBayes block summarizing the
partitions in the resulting matrix. Merged matrix can be mixed
(DNA, protein, and morphological) using MrBayes's modified
NEXUS format.
- Managing matrices and files
- fuse
matrices into a single matrix, even if matrices
refer to different taxon names (e.g., sequences of different
genes), using a table
of correspondences among taxa blocks
- archiving
of taxon names (permitting alternative names
for taxa in a file)
- include
taxa from NEXUS, NBRF or FASTA files into current file,
e.g. to add newly acquired sequences to an existing matrix
- import/export Phylip tree files
- Randomization (reshuffling) of matrices
can now respect taxa partitions; also can reshuffle within taxa
- PDF file saving
of graphics of trees and other windows
- Colors of character tracings can
now be changed
- Simplified installation process
for Windows computers
- A complete
list of standard modules has been added to the manual.
Bug Fixes
- ()
In previous versions of Mesquite, simulations and randomizations
of trees and characters used a shortcut to go to the n'th replicate,
cycling the (pseudo-)random number generator n times before
entering the replicate. This caused adjacent replicates to be
not entirely independent. In most cases we expect this bug would
not have had a biasing effect if sample sizes were large. This
bug has been fixed by adding 1 to the random number seed after
cycling n times to generate independent random number sequences
for each replicate.
- Various bugs and inefficiencies
affect memory use have been solved in file reading, tree display
and printing
- There were some copy/paste bugs,
especially under Windows, that have been fixed
- Swapping in heuristic tree search
now retains only unique trees
- NEXUS files with high-ASCII characters
(e.g., letters with accents) are better handled
- Various other minor bugs and inefficiencies
have been fixed
- build h61 (released 23 May 2006)
has a few extra bugs fixed: (1) Categorical likelihood calculations
no longer complain if root has zero length; (2) Bug in Java
1.5 on Mac OS X caused list dialogs not to respond properly
to clicking "Show Secondary Choices"; workaround implemented;
(3) Incompatibility with PDAP file exporter solved; (4) directory
choice dialog under Mac OS X improved.
Version 1.06
New Features
- The window showing annotatons for taxa, characters and cells
of a matrix has been integrated into the Character Matrix Editor
as a panel within that window. This makes it easier to annotate
without having to manipulate two windows. This panel can be
requested by the Show Annotations Panel item
in the Matrix menu, or by touching on the small button (
 )
near the lower left of the Matrix Editor (beneath the taxon
names). The pencil tool ( )
has disappeared. Another change is that images in annotations
can now be remote (i.e. URL's). Also, the annotations panel
is also available in the List of Taxa window (for annotating
taxa) and the List of Characters window (for annotating characters).
In the list windows, the panel can be requested by touching
on the small button ()
at lower left of the table.
- Annotations can now be attached to the states of a categorical
character, within the State Names Editor (available from the
Matrix menu of the Character Matrix Editor). The annotations
can be viewed by choosing Show State Annotations
in the State_Names menu or by touching on the small button at
lower left of the table ().
- A legend is now available for the colors in a Character Matrix
Editor. The legend can be shown by selecting Show Color
Legend in the Matrix menu, or by touching the small
button(
 )
at lower left of the Matrix Editor (beneath the taxon names).
If you double click on a color in the matrix, the editor will
move to a cell with that color.
- There are buttons at the lower left of the Character Matrix
Editor to open the List of Characters window (
 )
and the List of Taxa window ( ).
Reciprocally, the List of Characters window has a button ( )
to show the Character Matrix Editor.
- Character matrices can now be viewed in a "Birds-eye
view" with very narrow columns.
- There is a State Names Strip available for
categorical matrices under the Matrix menu. This shows a strip
along the bottom of the matrix in which the names of characters
and states are shown. Similarly when Bird's Eye View is used
for a matrix, a Bird's Eye Closeup Strip appears
below the matrix so that you can more clearly see the states
of the characters of focus. Both of these strips can be made
higher by going to their upper edges, clicking and dragging
upward.
- You can now view a character matrix in more than one window,
by requesting an Extra Matrix Editor from the
Characters menu. This is useful, for instance, if you want to
have the cells of the two matrices colored differently.
- The order of characters can now
be managed through stored character orders. A Default
Order of characters is automatically stored as characters
are created. This default order is visible in the List of Characters
window as the first column. If the order of characters is accidentally
or intentionally changed (for instance using the Sort
Tool) then the default order can be recovered easily
by applying the Sort Tool to the Default Order column of the
List of Characters window. In addition, you can store and reload
other character orderings by showing the Stored Character
Order in the List of Characters window (Columns menu).
You can store the current ordering and recover previously stored
orderings using the drop down menu from the Stored Order column
heading. When you do this, you will see the numbers in the column
change, but the characters will not immediately be reordered.
The reason for this is that what is stored and displayed is
simply a number for each character marking its position. It
is up to you to use that ordering to resort the characters.
Thus, to recover a former sequence of characters, load the stored
ordering, then touch on its column with the sort tool.
- There are new facilities for searching
for sequences in molecular data files. Thes are found
in the Edit menu for editor windows of molecular sequence data.
These are:
- Copy Sequence
(at bottom of Edit menu): This copies the selected cells
of the matrix into the computer's clipboard as a sequence.
That is, whereas the standard Copy would place into the
clipboard selected pieces of the matrix in tab-delimited
text format (e.g., if the sequence AATCA is selected, "A-tab-A-tab-T-tab-C-tab-A"
would be copied), this modified Copy Sequence command does
not include tabs (thus, "AATCA" would be copied).
This style of copying is useful when interacting with programs
like Sequencher (TM).
For instance, if you want to find a piece of sequence in
a matrix in Mesquite within a chromatogram viewer of Sequencher,
do the following: select the sequence in Mesquite, choose
Copy Sequence, then go to Sequencher, select Find Bases,
and paste the sequence as the search string.
- Maintain Target Match
(in the Find Sequence submenu of the Edit menu): This highlights
and maintains highlighted the first occurence of a given
sequence in a given taxon. First, you are asked which taxon
to search in. Then, it displays a panel like this:
underneath the matrix. The first button (red X) is to close
the panel; the second pauses the search; the third allows
you to select another taxon as your focus. If you type a
sequence into the text area, the matching sequence (if any)
will be highlighted in the matrix. Mesquite is constantly
monitoring this text, and so you don't need to give any
command to find again if you change the text. This is useful
if working with a program like Sequencher. If you see a
stretch of sequence while viewing chromatograms that you'd
like to find in the matrix in Mesquite, type in the sequence
into the text box and you will quickly be taken to it in
the taxon.
- Maintain Clipboard Match
(in the Find Sequence submenu of the Edit menu): This is
similar to Maintain Target Match, except that it obtains
the search string not from the text area but from the clipboard.
If the clipboard changes, the function will automatically
find the sequence again in the matrix. This is useful if
working with a program like Sequencher. If you turn on Maintain
Clipboard Match, then you can copy stretches of a sequence
within Sequencher, and Mesquite will automatically highlight
it, without your having to return to Mesquite or give any
other command to it. (Mesquite is constantly monitoring
the clipboard to see if it changes).
- There is an optional system for
assigning and maintaining ID strings to uniquely identify taxa
and characters. This can be turned on via the menu items Invent
Unique IDs for New Taxa and Invent Unique IDs
for New Characters in the Defaults menu of the Log
window. When enabled, newly created characters and taxa are
automatically assigned IDs which are maintained even if the
taxa or characters are rearranged through reordering, additions
or deletions. These unique IDs are little-used in Mesquite at
present, but eventually will permit users to reassociate taxa
and characters from different files even after names have been
changed (for instance, to re-read a tree file saved with different
taxon names). The IDs will also be used in database connectivity
and collaborative features. To assign IDs to already-created
taxa or characters, go to the List of Taxa or List of Characters
windows and choose Unique IDs from the Columns
menu. IDs can be generated using the drop down menu from the
column's header. The one disadvantage of assigning IDs is that
the IDs are stored in the Taxa block and Characters block of
the NEXUS file, which causes some programs to refuse to read
the file. (PAUP* by default will refuse to read these files,
but you can request that PAUP* ignore these extra commands by
the command "set errorstop = no;" which is available in its
Options Menu, Warnings and Errors.)
- There has been a shift in NEXUS file format for annotations
— the AN command is now used. This, unfortunately, produces
files that are not readable by version 1.05. The reason for
this change is to minimize text in the file when dealing with
a large matrix with thousands of annotations.
- Pairwise comparisons were previously forced
to find all possible pairs. Users can now impose a limit so
that only the first n pairs are found
and considered. This is important if there are too many pairs
to efficiently consider.
- Support for the new Cartographer
package.
- Groups of characters and taxa can now be assigned grayscale
values instead of just colors.
- Rarefy Tree (a random tree modifier) now uses taxon selection.
If some of the taxa are selected, than Rarefy Tree chooses which
taxa to exclude randomly only from among the selected taxa.
- There is a new option to export files, Old-Fashioned
NEXUS, for use by programs that understand only a small
subset of the NEXUS format such as MrBayes.
- Trees can be drawn in "Eurogram"
style by using Square Tree and selecting the menu item Cut
Corners.
- The Developer's
documentation has been updated
Bug fixes
- ()
Fixed bug that caused labels for internal nodes (including Bayesian
support values) to be scrambled if the branch interchange or
reroot tool was used.
- ()
The PAUPConstraintTestParsimony template included with Batch
Architecture for doing the test of monophyly has one of the
searches incorrectly searching for the optimal trees with the
converse of the constraint rather than the constraint
itself; this has been fixed.
- ()
Fixed bug that caused tree statistics to be displayed incorrectly
in columns in the List of Trees window if trees were deleted.
- Fixed bug in dialog boxes with lists from which to choose
that made them unusable under Mac OS X 10.4.
- Fixed bug that prevented use of high ASCII characters in NEXUS
files (e.g. "é", "ü", and so
on).
- Fixed bug that greatly slowed file reading under Java 1.5
- Fixed various issues to improve performance under Mac OS X
Java 1.4. (Some windows are still misdrawn, but that is due
to bugs in Mac OS X.)
- Fixed other minor bugs (including, but not only: table copy/paste
under Windows; use of arrow keys in matrix under Windows; retention
of full character information in concatenated matrices; stepmatrix
reading when character matrix ambiguous; fill tool with gaps)
Version 1.05
New Features
- added a new Find feature (in the Edit menu) that allows you
to search for sequences within DNA data matrices. The first
example of a module doing this is "Matching Sequence"
which finds sequences matching a target sequence the user enters.
The search allows a certain number of mismatches. Optionally,
it can search for the reverse, complement and reverse complement
of the target sequence.
- added a new item in the Select menu of Character Matrix Editor,
"Select Lowercase Ends". This selects
the ends of a DNA sequence up to the first block of N sites
in a row with uppercase symbols. The user chooses N. Assuming
that lowercase letters are used for less certain base calls,
this can be used to select and then trim (by painting with gaps)
poorly-sequenced terminal regions.
- the Concatenate Matrices command will now
allow molecular sequence data to be concatenated to the end
of a matrix of standard categorical characters.
- in the Mac OS X download, added an executable to run Mesquite
under Java 1.4. This may be less stable than the standard version
running under Java 1.3.1. The Java 1.4 version should be used
only if you have Java 1.4.2 update 2 or later on Mac OS X. (The
problems with early versions of Java 1.4 appear to be restricted
to Mac OS X.)
Bug Fixes
- ()
fixed bug in the Brownian motion simulations of continuous character
evolution. Brownian motion puts more change on long branches,
less change on short branches. The bug in the simulations caused
this contrast between long and short branches to be heightened
— long branches had even more change relative to short
branches than they should have. (Inadvertantly the effect of
branch length had been squared in the calculation.)
Version 1.04
New Features
- added Utilities submenu to Tree menu of Tree Window. Currently
there is only one utility, E-mail Tree, which
prepares a text image of the tree and puts it in the body of
an e-mail message requested using a web browser's mailto: command.
Bug fixes
- ()
fixed bug in two tree simulators, Uniform speciation
(Yule) and Uniform Speciation with Sampling.
In versions 1.03 and before, these failed to randomize tip labels,
resulting in trees whose shapes matches the model's expectations,
but whose particular relationships did not. Thus, in previous
versions the taxa were added one at a time to a tree growing
through time. For example, if there are 6 taxa, then the divergence
between the 6th taxon and its sister species was always the
most recent divergence in the tree. While this would not affect
questions about general tree symmetry (e.g. "What is the
distribution of tree asymmetries under uniform speciation?")
or branch length distributions, it would affect questions about
similarity between an observed tree and those expected under
a null model of speciation.
- fixed bug that caused a crash when characters were deleted
from a matrix which had annotations, images or colors attached
to the cells but which didn't have footnotes.
- fixed a few minor bugs.
Version 1.03
New Features
- added automatic file backup, which retains
copies of previously saved versions of files. This feature was
added to allow the user to return to a previous version in case
of errors or other problems. To enable this, in the Log window
or Projects & Files window select Automatic NEXUS backup
in the Defaults menu. Indicate the number of backups to be saved.
If you enter, for instance, 3, then when you save a file named
"myfile", the previous version of the file (if any)
will be retained and named myfileBKP1. It will be in the same
directory as myfile. The version before that will be called
myfileBKP2, the version before that myfileBKP3. When you save
the files again, the file ending in BKP3 will be deleted, BKP2
will be named BKP3, BKP1 renamed BKP2, and the newest backup
will be called myfileBKP1.
- enabled grayscale shading in Trace Character
History for continuous-valued characters. This can be turned
on in the Trace menu by selecting the "Use Grayscale If
Continuous" menu item.
- Matrices can be compared by selecting cells that differ.
To request this, choose "Select by Matrix Comparison"
from the Select menu of the character matrix editor. This is
similar to the Utility "Compare matrices", but instead
of giving a textual summary of the differences between two matrices,
"Select by Matrix Comparison" highlights those cells
of the matrix that differ.
Bug fixes
- fixed a bug that caused NOTES blocks to be written in duplicate
form into all linked files (opened via Link File...) if all
files were saved. This bug was generally innocent except if
a linked file was subsequently unlinked from the main data file.
This would leave the linked file with a relictual copy of the
NOTES block. If the main data file was then edited (e.g., footnotes
changed, characters deleted or moved), and then the secondary
file was later linked again to the main data file, its relictual
(and now out of date) NOTES block could override the current
NOTES block in the main file. This could cause footnotes in
the main data file to revert to former ones, or to be applied
to inappropriate characters or taxa. In addition to fixing this
bug, we have added a warning that is given if Mesquite detects
that a NOTES block is being overridden. If you suspect that
your footnotes may have been corrupted by this bug, please contact
us for assistance.
- fixed a bug in Trees Directly from File that gave spurious
warnings.
- fixed various other small bugs
Version 1.02
New Features
- multiple notes with images can now be associated
with taxa, characters and cells of character matrices. Prior
to 1.02 a single image could be associated with each taxon,
or a single footnote could be associated with each taxon, character
or cell of the matrix. Now, multiple notes can be attached to
each taxon, character, or cell of a matrix. Each note can have
a comment, a reference, an image, and an author. Labels can
be attached to the image. To use this, select the notes
tool ()
in the Character Matrix Editor. If you touch on a cell, a Annotations
window will appear with the notes for that cell, if any.
Buttons in this window can be used to delete and add notes and
images, and to make additional notes windows. Behavior of the
window can be modified using its Annotations menu. NOTE: the
single footnote system still exists and is independent of this
multiple notes system.
- MRP matrices can be generated from a set
of trees, for use by Matrix Representation with Parsimony analyses
to construct supertrees. To use this, select MRP Matrices from
trees as a source of matrices (e.g., under Characters menu,
Make New Matrix submenu, Other Choices).
- a search facility has been added to the Character
Matrix editor and to the List windows. The cell of the matrix
found by the search is highlighted and made visible. To
use this, select Find String or Find Footnote from the
Edit menu. Find String searches through the row names first
(e.g., taxon names for the Character Matrix Editor), then the
column names (e.g., character names), then the internal cells
of the table (e.g., the character state codings). Find Footnote
searches the footnotes of the Character Matrix and highlights
cells containing the requested text in their footnotes. You
can also search for stretches of sequences or state distributions
matching the currently selected stretch using Select Same Sequence
and Select Same Distribution of the Select menu of the Character
Matrix Editor.
- More choices were added for coloring character matrices.
Both text and background of a cell can now be colored according
to properties of the cell. A paintbrush tool (
 )
was added to allow arbitrary coloring of cells of a character
matrix.
- taxa and characters can be moved by a menu
item "Move Selected [taxa or characters]To..." in
the List of Taxa and List of Characters windows. Prior to 1.02
taxa and characters could be moved by clicking and dragging,
but using this it was difficult to move them long distances
through a matrix. To use this, select the taxa or characters
to be moved in their respective List window, and select Move
Selected from the List menu.
- Taxon selecting in the tree window has been modified, with
a new taxon select tool (
 ).
- Utilities to modify taxon names have been
enhanced and added to the Character Matrix Editor in the Taxon
Names submenu of the Matrix menu.
- Added import/export of FASTA files
Bug Fixes
- ()
fixed calculation of Shared Partitions, which counted too many
shared partitions between trees in some circumstances if the
trees had different terminal taxa included
- resolved conflict in interpetation of tree descriptions that
would result in Mesquite's interpreting trees from MrBayes and
other programs as having reticulations whenever labels for internal
nodes were duplicated. Mesquite now interprets these labels
as cosmetic only, and hence not as indicating reticulations
- fixed bug involving the State Names Editor, by which a character
might mistakenly receive a name duplicated from another character
- fixed bug in moving taxa or characters by dragging past the
end of the matrix
- fixed bug in recoding characters by which state names and
footnotes would not be adjusted
- fixed various other small bugs
Version 1.01
New Features
- added on-the-fly filtering and transforming of trees.
Thus, trees being supplied from a file or simulation can be
filtered according to various criteria or can be transformed
with the available tree or branch length alterers. To use
this, whenever a source of trees is needed, choose first
"Filter Trees from Other Source" (if you want to select
trees satisfying a criterion; more details here)
or "Transform Trees from Other Source" (if you want
to transform trees, e.g. reroot, scale branch lengths, etc.;
more details here), then
indicate what source of trees is to be filtered or transformed,
and what filter or transformation is to be used.
- added module to read trees directly from a file
one at a time; allows processing of much larger tree files (>50,000
trees). To use this, request "Trees Directly from
File" as your source of trees for charts and other calculations.
You can also build a tree block from a sample of trees from
a large file by selecting "Include partial contents"
from the "Get File With Trees" submenu of the Taxa&Trees
menu.
- available tree and branch length alterers can be applied to
all trees in a tree block. To use this, go to the Utilities
submenu of the List menu of the List of Trees window.
- added a new tree alterer: outgroup rerooting
of trees. To reroot trees using a selected set of outgroups,
you first must select the outgroup taxa, then, choose the menu
item "Root tree with selected taxa as outgroup", for
instance in the Alter/Transform Tree submenu of the Tree menu
of the tree window. More details are given on the page on trees.
- added integrated BLAST search. To use
this, go to a character matrix editor window showing a
molecular matrix. Select one or more stretches of sequence and
choose BLAST Search from the Search submenu of the Matrix menu.
Mesquite will then send a request to your web browser to perform
a BLAST search at NCBI's BLAST server, one request for each
of the selected stretches of sequence. The results will be available
in your web browser.
- added -w (windowless) and -b (backgroundable)
flags for running long Mesquite jobs in the background in Unix
from the command line
- improved user interface of tables (List Windows, Character
Matrix Editor), including navigation with arrow keys and more
informative explanations and footnotes
- Added autotab to the Character Matrix Editor.
With this, the edited cell shifts automatically to the right
or downward when a state is entered. To use this, select
the Select and Type tool (with the letter "K") and
use its drop down menu to indicate autotab right or autotab
down. Autotab functions only when the Select and Type tool is
the active tool.
- selection of taxa, characters and trees is now remembered
when files are saved
- calculations of charts and Trace Character Over Trees can
now be stopped part way to show partial results
Bug fixes
- fixed tree printing bug by which node numbers and branch lengths
always appeared
- fixed bug by which pressing the Delete key while editing in
a table would be misinterpreted as a requesting to delete a
character, taxon or other object
- fixed bugs in NEXUS file reading and writing (concerning SYMBOLS
list, blank character names, tokenization with tab or square
brackets, near-duplicate taxa blocks in same file)
- fixed bugs in State Names Editor and Character Matrix Editor
concerning footnote display and editing
- fixed bug with Mac OS X 10.3 by which dialog boxes with lists
always chose first item (this was not a bug in Mesquite, but
rather in Mac OS X, but we have found a workaround)
- fixed bug with Windows by which pie diagrams would be misdrawn
as entirely black for traced characters in the Balls and Sticks
tree drawing mode (this was not a bug in Mesquite, but rather
in Windows/Java, but we have found a workaround)
- several other smaller, less notable bugs were also fixed
|
