v4.397
1)Ensures that with ONE CLICK you can turn on or off the highlighting (i.e. back-gating) of MULTIPLE subsets from the Phenogram window
2)Fixes bugs with latest Pbe (probability bin embedding) based on tSNE which is launchable from N2D views of MDS and Phenogram
v4.396
1)Ensures Pbe feature’s window comes up more smoothly from MDS display
2)Support Pbe feature in Phenogram along with easy highlighting toggling
v4.395
1)Fix 17 month old problem with tSNE where the scatter variables are not “normalized” using the Log10 idea of Wayne Moore
2)Make it easy for user to toggle highlighting ON and OFF by clicking on the symbols in the MDS display
v4.391
Improves Pbe
v4.393
Introduces new feature “Probability bin embedding” which is tSNE with probability binning which makes subset separation less crowded and makes tSNE faster and makes it reproducible.
v4.392
Introduces a new feature asked for at Dec 17, 2018 lab meeting for showing degree of match similarity with heat map colors based on earth-moverâs distanceâ¦this helps see quickly degree of difference in one glance. One activates this from the EMD toolbar button on the MDS display.
v4.391
Adds to heat map:
1) Legend to describe marker expression
2) Column for frequency
v4.390
Provides heat map column headers rotated 45 degrees in major commercial browsers (google chrome, safari and edge)
v4.389
Fixes bugs related to the new GUI for Epp and MDS and phenogram
v4.388
Makes heatmap tool tips better
v4.387
1)Support vertical labels for markers in heat map
2)Don’t show scatter in heat map
3)Handle larger phenograms >25 subsets
v4.386
Deals with many issues discovered on Microsoft Windows high definition monitors with new pipeline MDS and Phenogram
v4.384
1)Improve phenogram so tips work better and png files are created with progreess bar
2)Improve heatmaps to :
a)use select-able color schemas
b)use fire color scehma by default
c)not show underline in links
d)be shown in browser
v4.383
Bug-fixes and stabilizes the phenogram view
v4.382
1)Allow user to see multiple gating sequences in single web page from MDS window or Phenogram wiondow
2)Ensure order of branches or leaf in phenogram follows the EXACT phenogram order or the reverse order
v4.381
Fixes surprising problem with MatLab compiler where it does case INSENSITIVE comparisons of module names unlike the full MatLab environment. Thus the build in heatmap function gets confused with AutoGate’s HeatMap.m module. WOW. YIKES. SIGH.
v4.380
Makes the Phenogram view as functional as MDS view
v4.379
Upgrades median marker expression to phenogram view when user clicks N2D toolbar button. Subset rows in heatmap are ordered by same order as phenogram. User can alter subset order.
v4.378
Introduces median marker expression to phenogram view when user clicks N2D toolbar button. Subset rows in heatmap are ordered by same order as phenogram. User can alter subset order.
v4.377
Version 2 of feature “Median marker expression” for HiD Subset View (MDS)
v4.376
Version 1 of median marker expression heat map which CDW partricipants deemed essential to auto gating flow informatics review.
v4.375
Fixes recent bugs and adds ability to see marker expression on ND Subset View (MDS)
Possible fixes for Lisa ‘s experiment
v4.373
Makes cosmetic improvements for slides at CDW
v4.372
1)Support QF + CityBlock for phenogram
2)Support debug dump of score matrix
3)Reduce scatter sdu max to 1
4)Clean up the display of parameters on 1D PathFinder
5)Add match info to legend on MDS display if not 1×1
v4.371
1)Wording changes such as “dendrogram” to “phenogram”
2)mds display (HiD subset view) makes important match information more quickly available on context (right-click) menu and tool tip hover hours. For example it adds info such as âMatch: 1×3 subset, 12,302 x 14,987 cellsâ.
3) AutoGate prevents HiD subset display and Phenogram display when user chooses only one FCS parameter
4)Improve left side of AutoSinglet’s polygon gate when using Wayne’s density algorithm.
a)Improve symbol namesof programming to clarify latest design intentions.
b)Optimize output of merging to only deal with best results for collecting all best matches
c)When optimizing merging by F measure do not always take top F measure if other scores are close. Be tolerant of F measures that are close to each other.
d)Corrects the scaling coefficient of QF+euclidean distance for dendrograms to (maxDistance/minQf)^(-1).
v4.369
consistently computes the QF + euclidean distance described by Guenther.
v4.368
adds a new dendrogram distance: QF + euclidean
v 4.367
1)Improve wording and text positions in dendrogram1)Improve wording and text positions in dendrogram
2)Allow “all subset” remember-my-answer for acceptable subset overlap for QF + EPP at session and non session levels.
v4.366
Fixes several bugs found by Wayne trying to improve EPP
sortDir NEEDS to inform the context invoking it what the original unsorted index is for each sorted description
1)Save probability bin size for display at top of window title
2)Support F measure on probability bin merger
3)for quality of match test use 2 deviation units apart for scatter versus 4 for staining. User can opt out of this bullshit
a)Save probability bin size for display at top of window title
b)Support F measure on probability bin merger
c)for quality of match test use 2 deviation units apart for scatter versus 4 for staining. User can opt out of this bullshit
v4.363
Adds new feature of COMPARING and finding differences between HiDSubsetView windows. User simply drags one such window on to anojther.
v4.362
1) Restore 4sd check for merge candidates
2) Improve title info for gating sequence choice window that is called from MDS window
3)If ClusterColor scheme is used with 2D look ahead then always use color scheme even when the clusters are the same as the parent subset.
v4.361
Contains many improvements made while in Italy. The main ones are
1) Merge strategy alternatives where candidates can be not only the best matches specially but among the best OVERALL matches
2) F measure logic for sanity testing, for optimization and for an alternate Match strategy where goodness of match is not based on QF score but on F measure only.
v4.361
Contains many improvements made while in Italy. The main ones are
1) Merge strategy alternatives where candidates can be not only the best matches specially but among the best OVERALL matches
2) F measure logic for sanity testing, for optimization and for an alternate Match strategy where goodness of match is not based on QF score but on F measure only.
v4.360
Frequency basis in QF tree should be same as MDS display (HiD Subset view) if invoked from MDS display
v4.359
1) fix agi getting too big with images of deleted gates
2)font preferences for MDS (HiD subset) display
v4.358
1)Make QF tree window consistent with other AG windows for right-click, hovering and genie search (60% done)
2)Put F measures in MDS display
3)Use F measures to optimize merging with QF match
4)Remove Darya’s prior required optimizations for QF match since they were not safe
5)Work around oddities in FlowJo 10 export of FCS file
6)Allow QF tree to be exported to GatingTree as Boolean gates
7)Support font preferences on QF tree
Improve click GUI on MDS screen so that critical diagnostic options can be reached more soon
v4.356
1)Nicer display of gating tree sequence window from MDS window
2)Allow QF score display on QF tree by user preference
Improve click GUI on MDS screen so that critical diagnostic options can be reached more soon
Improved cosmetics for HiD Subset Viewer
Memory of answer to hi res question was buggy
Fix exit question
v4.352
improves messaging of HiD Subset Viewer for hovering and clicking. More consistent, more netat.
v4.351
Fixes bug where if you pick an EPP gate first and then a non EPP gate then on the 1DPathFinder display of 2 gates has the +/- deviation information and units is INVERTED. Quality of match which is done on EPP gateâs median/deviations is instead reported on the screen as if it had being done on non EPP median/deviations.WRONG.So this question now occurs if the user picks EPP first and then non EPP. IF they answer NO then the correct reporting NOW occurs ⦠but is probably not what the user should do since the conventional standard to compare against is known gates and NOT automatic gates.
v4.352
Improves messaging of HiD Subset Viewer for hovering and clicking. More consistent, more netat.
v4.351
fixes bug where if you pick an EPP gate first and then a non EPP gate then on the 1DPathFinder display of 2 gates has the +/- deviation information and units is INVERTED. Quality of match which is done on EPP gateâs median/deviations is instead reported on the screen as if it had being done on non EPP median/deviations.WRONG.So this question now occurs if the user picks EPP first and then non EPP. IF they answer NO then the correct reporting NOW occurs ⦠but is probably not what the user should do since the conventional standard to compare against is known gates and NOT automatic gates.
v4.350
1)Ensure that experiment properties take precedence OVER global properties. SIGH
2)Simplify saving of analysis to be used as drag&drop window when changing compensation or biexponential (logicle)
v4.349
1)fix 5 year old problem in single-stained control display which users invoke from PlotEditor. The compensating out of the single stain into unstained channels had a programming bug in matrix alignment
1)Support limiting FCS parameters for 1D pathfinder on QF tree to only those used for the tree
2)Show mean/median symbol on 1D pathfinder next to arrow and to text to clarify this is NOT the peak
3)Allow user to set the visual min or max of the X and Y scale for CYTOF, Synthetic and FACS data
4)Tweak the min end of a visual scale to avoid hiding the cluster borders
5)Logicle scale improvements
a)Position the min scale for clustering based on the .5% percentile of the negative of all samples for the same stain set
b)Allow the user to define the
i)# of events to look at in EVERY sample for the same stain set
ii)minimum # of events in sample to make it worthy of democratic W consideration
iii)Percentile for the W (default=5%)
iv)Percentile for the minimum scale for DBM clustering (default=0.5%)
6)Clean up code for managing visual plot scale versus cluster scale to make it more understandable
7)Support CYTOF data sets better to
a)Avoid events that are OFF scale
b)Provide more narrow scales where the tickmarks and labels were INCORRECT.
c)Ensure axis choices for the PlotEditor do not include uncompensated data for CYTOF or any data set declared as already compensated
8)Accelerate the processing of democratic W for both CYTOF and fluorescent cytometry so that the onscale calculation is done only ONCE.
9)Support “extension ranges” for fcs column ids. Previously only 1 range was available signifying different things for FSC versus SSC versus stains. This is preparation for supporting linear scale on stains and log scale on stains and scatter
FIx to retain fcs file reassignments for autocomp
v4.346
1)Improve probability contour display using more narrow bandwidth for density. Wayne’s idea. Improves pseudo color too. Users have noticed that contour lines some times go over areas with no data. They will notice this less.
2)Help user not lose prior analysis if setup changes compensation and help them open a concurrent window of prior analysis in order to rebuild any gates that becamse locked and read-only.
v4.345
Fixes needed by infomratics reearchers (not biologists uses)
1)Improve Auto Singlet logic …. it is close to as good as possible.
2)Improve coefficient of separation
3)Allow user to see counts for things in the 2 lists of Bangalore’s Orderer module. Users complain about always having to count parameters when designing EPP inputs.
4)Fix latest problems in tex processing in ploteditor titles
v4.344
1)Make sure ALL negative data is considered in the clustering. User must reset experiment to get this improvement.
2)Bug fixes and improvements to applying Boolean gates between samples.
3)Other fixes to FATAL bugs caused by last 2-8 days of intense progamming.
4)some improvements to the Auto-Singlets by Wayne
Fix infinite loop with Microsoft windows due to their NOT following UNIX file naming standars like MAC does.
v4.342
1)MAJOR new feature for QF match where the user can DE-SELECT leaf gates that have duplicate events in them. When running QF Match
2)Fix to A serious visual bug caused by Stephen in last 8 days where the scale for logicle axes does not go down to 0 if you select gates high above zero and subset. This does NOT affect semi-automatic gating or gating
3)Fix to older serious bugs in agi file import/export where the
a) users can lose their analyses because export first does no work.
b) user can NOT export only 1 analysis (the primary one in .autoGate sub folder)
c) user loses pre-existing analyses not FOUND in the agi file
v4.341
1)Ensures Epp statistics are not lost from epp top and that the same epp top does not get the wrong configuration.
2)Increase accuracy and usability (e.g. cancel button) of keep disk space low in cached files and old downloaded files
3)New feature “Open secondary analysis” allows multiple GatingTree windows+analyses to be open for the same experiment/samples
v4.340
1)Fixes exception during Epp
2)Makes auto complete for XY axis choices better by allowing all items to be seen and escape key always losing
3)Has restrict button on PlotEditor window inform user about SIZE of restriction in % terms
The no match triangle symbol is being asked to look like any other non match making it harder to judge deegree of no match.
v4.335
A) Provide better support with CYTOF Logicle W and negative data better
B) Steve DeRosa’s assay required that we stretch the visual axes so that the 5% of the negative data is not hidden
C) Do not popup X/Y choice list UNLESS user clicks on the axis. Otherwise bring up auto-complete list.
v4.332
1) Ensure that user can see negative data below the logicle W point since both Darya and I discovered recently the logicle value can be < 0 and > 1. To allow DBM clustering to ALSO capture data below the logicle W you must change Density.m static method DbmVersion to return ‘v4’
2)Adjust handling of logicle T value for Darya’s CYTOF sample
3)Ensure large matrices can scroll and be seen
4)For boolean gates display parent set as most common ancestor # of events and NOT sample # of events. Show % too
v4.331
1)Ensure special gate icons (Epp, boolean, flowJo) are consistently embedded
2)Ensure boolean gates are created under most appropriate parent subset as is seen in FlowJo 9.x with Steve DeRosa’s workspaces.
3)Comment out fprintf DEBUG statement made a week ago which causes FATAL exception when creating publication images from GatingTree window
v4.330
1)Fix bug with compenstion matrix loading
2)Improve display of democratic w
3)ensure democratic w only computed ONCE
4)improve display and comsetics for boolean gate
v4.329
A) When running Epp AG allows user to
1)FIRST create auto-singlets and auto-live gate if necessary
2)automatically create an FCS sample with gating info when done
3)specify running the same Epp configuration on one or more other samples in the equivalent position
B) AG provides ability to find dendrogram/phylogenetic feature with genie search on GatingTree or HiDSubsetView windows
C) When showing dendrogram/phylogentic/QF tree this version of AG NOW allows user to click on ANY node and bring up a high quality matlab fig for 1D pathfinder. If the node is for multiple gates a BooleanGate is created to allow PathFinder to see measurement spread for combination of gates. If the node is for a single gate AG adds 3 toolbar items that when clicked allow the user to view associated gate’s
1)comic strip of gating sequence
2)PlotEditor window
3)Position in GatingTree and GatingTreeTable windows
v4.328
1)New feature to support export of Epp gates to FlowJo 9.x. If user selects fcs file type when saving samples and Final leaf gates … then they are allowed to convert Gate ID values into values that are distinguishable on the FlowJo 9.x logicle scale.
2)Run Wayne Moore’s COS (coefficient of separation) on each Epp separation AFTER epp is done.
v4.327
1) Accelerate applying of gates via drag&drop or cut&paste by finding and skipping-OVER equivalent gates already made for the current cluster context (cluster setting + X/Y parameters)
2)Support applying Boolean gates correctly (observing analogous dependencies) when applying gates with drag&drop or cut&paste
3)Support applying Auto-Live and Auto-Singlet gates when applying gates with drag&drop or cut&paste
v4.326
Improves Auto live gate. It now
1)uses name “Auto Live”
2)forces VERY low cluster detail
3)Uses GetPeak at 14 widths like Singlets to take max live marker value rather than max
v4.325
saves user’s preference for # of parameter medians >= 3
Change word “cells” to events
v4.323
1)Ensure 1D PathFinder reports event counts
2)Improves 1D PathFinder display for synthetic data. Not perfect yet. Much better.
3) More options in controlling QF Match
4) For QF tree only you can choose debug output levels to be sent to browser
5) 2 new cutoff methods for singlet gates
v4.321
1)allow user to open PlotEditor on “related compensation” popup
2)fix bug with samples lacking FSC-W and having LESS THAN FIVE events!!!!
v4.320
1)Provide ability to hide unmatched subsets in HiD Subset View
2)For auto singlet gating by deviation units, do a least squares fit of the plot points computed using standard deviation units
3)Provide boolean gates by tree picks or by ANY Boolean expression of gate-IDs plus AND/OR/NOT.
4)Allow direct importing of FlowJo 9 exported gates with text #
5)Display info about special gates (Boolean + FlowJo 9 import) in both the GatingTree and PlotEditor windows
6)Provide a Coefficient of separation for Epp gates after they are created to aid in pruning gates
7)MUST warn about overlap with QF Match + Epp because phenotyping requires ZERO overlap. Every cell lands in only one final gate of a sequence.
8)For 1D PathFinder
a)If not matching use case ALWAYS show the +/- mad (robut deviation)
b)If matching then
i)Only show +/- mad only for non Epp gates and not for Epp gates
ii)Hide the deviation units difference for non Epp gates
iii)Give the user the ability to merge 2 or more matching Epp gates into 1 view
v4.319
1)Avoid exception with geomean with negative numbers
2)Allow stdev unit of 1.5 for computing Singlet gate
3)Confirm IF the user wishes TRULY to apply Epp gate sequences…and allow them to ONLY apply non Epp gate sequences
4)Allow user to cancel out of LONG applying of gating sequencs (Epp or non Epp)….
5)Report # of cells in manual gate
6)Allow user to cancel creating png files for 1D PathFinder
v4.318
1Accelerate QF computing for matching by eliminating unnecessary merges>3
2)Accelerate QF computing for Dendrogram by not recomputing identity matrix duplications
3)Improve name editing for user with QF match
for v 4.316
Optimization created HUGE bug
HiDSubsetView window is now more consistent with the conventions of AutoGateâs other main windows.
a)The options menu toolbar is the conventional wrench icon and not the 1D PathFinder button.
b)Moreover, at the bottom left of the window is the conventional genie that allows you to search functions for the current window.
v4.315
1)Lee’s new feature to highlight related QF matches
2)Improve name matching
3)Support DEBUGGING QF matrix
4)Improve gate editing logic from legend
v4.314
1)closes child windows associated with the parent window when the parent window closes (e.g.1DPathFinder child or QfTree child of HiDSubsetView parent)
2)Position 2 Qf tree windows visibly when invoked from HiDSubsetView
3)Allow user to REBUILD QfTree with new parameters when invoked from HiDSubsetView
4)Highlight split-matches and no matches so they are more visible than 1-to-1 matches
v 3.13
1)Support refresh of names if user chooses QF Tree and names are added.
2)Do not recalculate QF Tree every time.
3)Allow user to activate both QF trees
v4.311
provides first basic draft of QF tree using MatLab’s simple visualization with their Dendrogram command
v4.310
1)r2018a warns against functions name “count”
2)Cosmetic 1D PathFinder changes
v4.309
Makes the 1D PathFinder display better for 3 parameters like Darya’s recent Synthetic experiment whether the display is the png file saving with ‘Re-create….’ or svg graphics saving.
v4.307
Improves automatic singet and live gate
v4.306
New logic for guarding against user misjudgement of distances with HiD Subset View
1)add numbers to axes
2)make X and Y scale the same length
3) make aspect ratio square
v4.305
1)Improves vector graphics for matching 1D PathFinder images
2)Provides preliminary support for inputting more than medians into MDS for HiD Subset View
v4.303
1)Allow arrows to be turned OFF
2)Remember screen positions
3)Use matched names as labels
4)Allow editing of axis labels and plot titles
v4.302
2018a and JIDE (SAFE commits)
v4.301
1)Preliminary automatic Singlets gate feature from GatingTree and PlotEditor
2)Preliminary 2D PathFinder logic to support automatic Live gate
3)New feature for HiD Subset View that allows flexibility in frequency basis of gates
4)Allow HiD Subset View with no matching to be run on ONE pick in GatingTree and Gating TreeTable.
5)If GatingTreeTable HIDES the Epp branches and user picks Epp branch in GatingTree … allow HiD Subset View or QFMatch to run on that selection
If epp top gate of time and fsc-a has no clusters then make full scale rectangle manual gate
If NOT isdeployed (matlab compiled runtime) ask user if they wish to use JIDE table
Providing detail report when more than one instrument channels are used in the selected samples folder and managing post experiment separation actions
Use JIDE – Preference can be set from CytoGate.m static method
Support delay in starting the shoing of a tip
Contact us on To do screens and Stephen’s other changes
Fix for corrupting side-effects even though it only occurred when creating manual gates and then clicking in the data area outside the clusters overlapped by the gate
A bug fix to very recent refactoring of sample file creating.
v4.264
Upgrades workspaces mad BEFORE v4.264 to ensure safety of all samples having both “Event #” and “FSC-W” parameters regardless of prior data caching or logicle W caching
v4.263
1)Ensure all samples have “Event #” and “FSC-W” parameters regardless of prior data caching or logicle W caching
2) Allow enter key to cause filtering of found rows in search JTextField of PathFinder’s score/rank JTable
3)Put “Event #” parameter towards bottom of X/Y selection list
4)Allow user to search on all gates or non epp gate
5)Use pretty EPP icon to mark Epp gates instead of the string “*Epp”
6)Make sure restore button on histogram gate does not vanish
v4.260
Replaces Yair Altman’s treeTable for MatLab when sorting score rankings on plots in 2D PathFinder
v4.259
1) Deal with GatingTree property files that become too big when swollen by many Epp and 2D PathFinder operations: a)remove all cluster context orphans; b) remove all unneeded png files and tSNE_*.mat files.
2)Add Young David Parks ideas to have a FINAL epp gate with only the 1 side of the parent gate that is selected and THEN select the final 1 cluster with 0% cluster outliers.
v4.258
1)Improve the logic for balanced separation
2)Default minimum split cluster size is 75
3)1st draft of logic to cleanup orphan cluster contexts
4)Make highlighter gate identifiers consistently more descriptive using parent and sample information and size information
5)Accelerate F measure computing for highlighting webpage
v4.257
1)Make consistent the wording about best separatrix split option.
2)Fix bug where separatrix scores could no longer be viewed from the PlotEditor window
v4.256
1)Fix typo with ‘p’ character
2)Ensure Epp top gates that consolidate config settings for epp for a distinct set of cells are not duplicated
v4.255
1)Adds more fixes for Microsoft Windows High Definition problems
2)Support Epp feature/function where Epp settings and sub gates can be coordinated and remembered in a single gate under the gate chosen for Epp.
v4.254
Provides additional webpage that shows the best F measure for gates overlaid by highlighted gates. This speeds up assessment of the effectiveness of Epp input configurations.
v 4.253
1)Fix bad finite progress bar when removing gates
2)Fix infinite progress bar disappearing too soon when closing GatingTree window and savingProvide tooltip and high/3)low support for every key typed into number field for Epp or highlight webpage
4)Allow Epp user to configure the balanced % to determine the smallest portion of half that is considered “balanced”
v4.252
1)Fixes problem where after the first recoloring for the same gate on the same dialog window is not shown the next time you click the flashlight. Dialog must be closed.
2)Accelerates resolving of name duplicates which becomes slow after runnint Epp.
3)Does not suggest Epp names to user when user is naming a gate.
4)Show only one gate’s highlighting when showing highlighted gates on the highlighting webpage report.
v4.251
1)Fix seriously-stupid bug in backward compatibility for old % outlier calcuations
2)Improve opening/closing of large GatingTrees bloated by Epp
3)Improve sample descriptions for highlighting webpate by showing images in table
4)Improve transparency
5)Better wording for start/stop/recolor multiple gates
v4.250
1)Fixes new exception bugs in reset/restore/save analysis on File menu of GatingTree window. These were caused by optimizing the flashlight/highlighters in December 2017.
2)Continues fixing MS Windows pc problems with high definition monitors.
3)Makes “never reuse” the default mode for Epp
v4.249
1)If user chooses highlighting webpage where only images are chosen where highlighting meets the criteria THEN show all parent images for highlighting FROM and highlighting TO gates
2)Fix programming bugs with most equal/most balanced mode of GatingTree
Log usage even when Authentication is turned off
v4.247
Fixes a fatal exception with Epp that can occur with certain user preferences. This was caused by a cosmetic improvement 24 hours ago.
v4.246
Continues finding & fixing PC Windows high definition monitor problems while testing Epp
v4.245
1)Fixes Epp programming bug where the minimum size for clustered cells to split on was not working in projections that only had 2 clusters. Sigh.
2)ALWAYS show the gates being highlighted from at top of report
3)Support ONLY showing gate images that meet overlaying/overlaid % criteria
4) Provide menu items for “Highlight sample subsets” and “Stop highlighters”.
5) By default show non Epp gates in highlighter and require user to ask for “epp and non Epp”
v4.244
% cluster outlier inclusion was rough from the very beginning of this feature BECAUSE it was based on trying to optimize distance calculations using DBM’s unique grid bin locations for events instead of the eventâs exact locations. Since % outlier is subjective like manual gating this roughness was not considered to be more than a minor problem. This is too ROUGH or inapproximate for Epp which uses 100% of the nearest cluster outliers and thus it needed to be fixed.This creates a backward compatibility issue since any gates that used % outlier >0 before this fix in v4.244 can not be reproduced without the old rough logic. The more accurate % outlier calculation will not find these gates.So … Instead of locking these older gates and making them non useable to the suer, however, the old % outlier calculation continues to be supported for old gates but can not be used for a NEW gate.The user is warned and advised to upgrade the gates.
