***** 05-21-2015 ver. 2.0.0.33 bug fix: curve fitting (custom curves) since 2.0.0.31 ***** 02-26-2015 ver. 2.0.0.32 bug fix: reading ABF files ***** 07-29-2014 ver. 2.0.0.31 Multi-threaded MIL (with multiple data segments or list selections) ***** 07-28-2014 ver. 2.0.0.30 Warning when running MIL with channel count too high. ***** 03-31-2014 ver. 2.0.0.29 Clarification: events with duration of exactly Tdead are not dropped. ***** 03-14-2014 ver. 2.0.0.28 Added support for National Instruments USB hardware with differential-only inputs ***** 02-27-2014 ver. 2.0.0.23 Save Idealized (DWT): writes more decimal places, to work with high sampling rates ***** 11-26-2013 ver. 2.0.0.22 Fixed Scaling factor when reading Igor Binary (IBW) data ***** 04-26-2013 ver. 2.0.0.21 Supports QtiPlot built with Python 2.6 or 2.7 ***** 03-22-2013 ver. 2.0.0.20 Simulation: Dose-Response: added "Current Integral" to Results:Segments ***** 02-01-2013 ver. 2.0.0.19 correctly opens .dat files from PatchMaster ***** 12-18-2012 ver. 2.0.0.17 bug fix: saving and extracting idealized data, in locales using the decimal comma ***** 12-06-2012 ver. 2.0.0.16 Scripting support ***** 11-23-2012 ver. 2.0.0.15 New in MIL Results' Reports menu: Excel Histograms Data drawing fix (Graphics mode, no overlapping channels): don't draw data that's out of bounds ***** 08-02-2012 ver. 2.0.0.14 Put custom python scripts in Users\Public\QUB\PythonScripts (files here override other locations) ***** 05-17-2012 ver. 2.0.0.13 bug fix: Setting up new NI hardware ***** 02-26-2012 ver. 2.0.0.12 bug fix: Results and international date formats ***** 02-21-2012 ver. 2.0.0.11 Data keyboard shortcuts (e.g. 'J'oin) now act on last-clicked panel (lo-res or hi-res) Warning on cut/delete data ***** 01-25-2012 ver. 2.0.0.10 Changed Results:Select charts from log(base e) to log(base 10); If anyone needs log(base e), please tell us. ***** 11-30-2011 ver. 2.0.0.9 Experimental support for reading TacFit output table as an idealization; the table must be in a file with the extension .tacidl Adjusted action button font size Fixed "edit buttons" dialog for non-standard monitor DPI ***** 09-21-2011 ver. 2.0.0.8 Point at a window to select it (i.e. for keyboard input); no longer need to click in Mac Rates: improved "flat stimulus" warning ***** 09-09-2011 ver. 2.0.0.7 Model: Kinetic constraints: Generalized constraints now fully functional ***** 09-06-2011 ver. 2.0.0.6 bug fix: starting QuB by double-clicking a document ***** 08-18-2011 ver. 2.0.0.5 Idealization results: histograms of active A/D channel only ***** 08-09-2011 ver. 2.0.0.4 right-click in data to copy the tool-tip bug fix: Dose-Response simulation was missing final trace ***** 07-14-2011 ver. 2.0.0.3 bug fix: acquisition hardware setup ***** 06-23-2011 ver. 2.0.0.2 Opens Patchmaster .pul files (version 9). Thanks to Dr. Bernd Letz and HEKA. ***** 05-11-2011 ver. 2.0.0.1 bug fix: EVariantError in MacRates, since 2.0.0.0 ***** 05-05-2011 ver. 2.0.0.0 more stable eigenvectors for better LL convergence upgraded Python to 2.5 with numpy and scipy acquisition can be controlled by other programs - see http://www.qub.buffalo.edu/wiki/index.php/WM_QUB_ACQ_CONTROL - micromanager driver coming soon licensed under GPLv3 (see Program Files\QUB\LICENSE.txt) less clutter ***** 11-22-2010 ver. 1.5.0.39 bug fix: opening some ABF files bug fix: loop balance constraints with Ligand dependence ***** 11-03-2010 ver. 1.5.0.38 Idealization histogram pdf now uses "amp" and "sd" which are estimated without dead time Send to QtiPlot: fixed to work with single-user installations of Python 2.6 ***** 10-21-2010 ver. 1.5.0.37 Idealize, ChopIdl, Stat: results include 'interburst' -- time since last burst or start of data segment ***** 10-19-2010 ver. 1.5.0.36 Bkg Sub (Preprocessing): Remove periodic noise (e.g. 60 Hz) using a Kalman filter Stimulus (Preprocessing): Separate time constant for collapsing Model: negative conductance (pS) allowed Bug fix: simulating segments longer than 1 million points using a protocol Fully compatible with QtiPlot 0.9.8.3 (must install Python 2.6 and numpy) ***** 10-06-2010 ver. 1.5.0.35 Scripting enhancement: Results (Select) can be auto-curve-fit. ***** 09-28-2010 ver. 1.5.0.34 Send to QtiPlot: customize sub-menu by adding files under Program Files\QUB\QtiPlotScripts Bug fix: high-res highlighting when there is a stimulus variable ***** 09-13-2010 ver. 1.5.0.33 Send to QtiPlot: bug fixes and formatting improvements ***** 09-08-2010 ver. 1.5.0.32 Right-click any histogram or results-plot to "Send To ->" Notepad, QtiPlot, or Excel (QtiPlot requires customization -- please ask in the web forum) Bug fix: Break segment no longer corrupts idealization and baseline Bug fix: Break segment no longer corrupts text files Bug fix: Calibration Wizard was generating reciprocal of correct answer ***** 08-02-2010 ver. 1.5.0.31 Bug fix: saving models which were originally created in QUB Express ***** 07-26-2010 ver. 1.5.0.30 Results:Select menu: X Means web app Bug fix: Mac rates works when there are superfluous ScaleExp constraints ***** 06-09-2010 ver. 1.5.0.29 EditIdl: Join class ___ events to the prior event Model: Each rate can have a custom format. (The default format can still be overridden in Model Properties: Display.) For example: %f 0.123455667 %.3f 0.123 [%.3f]! [0.123]! %.3g 1.234E-1 see http://www.cplusplus.com/reference/clibrary/cstdio/printf/ Acquisition SShot: Resistance panel can compute MOhms = 1000 * mV / pA, where pA is read from an AI Channel, and mV can be read from a channel or entered as a constant. Acquisition SShot: Analog In channels display the mean of each buffer while acquiring bug fix: MIL "list index out of bounds [0]" on zero-event histograms ***** 05-24-2010 ver. 1.5.0.28 bug fix: EditIdl could create zero-length events ***** 05-19-2010 ver. 1.5.0.27 bug fix: decimal numbers in dialogs are read correctly , in regions with the decimal comma (where 1,5 * 2,0 = 3,0) (dialogs which show numbers using the decimal point (.) expect you to use it too; a few respect the comma) ***** 05-14-2010 ver. 1.5.0.26 Idl->Lst button makes a selection list from one idealized class bug fix: Simulate Dose-Response sometimes didn't work ***** 04-27-2010 ver. 1.5.0.25 bug fix: Data Overlap window, curve fitting AVerage or Sum ***** 03-09-2010 ver. 1.5.0.24 bug fix: avoids "floating point overflow" when results contain infinity or NaN ***** 03-08-2010 ver. 1.5.0.23 EditIdl: "Change class" and "Fill blanks" are much faster ***** 02-02-2010 ver. 1.5.0.22 X-Means option: "normalize each var..." - enabled: divides each var by mean; recommended for fitting multiple vars with different scale - disabled: calculates average squared distance (per point) in var units. ***** 02-01-2010 ver. 1.5.0.21 Save Idealized Data as ASCII: writes class amplitude instead of class number. Renamed Results:Select:"K-Means random" to "X-Means" Improved X-Means formula, see http://www.qub.buffalo.edu/wiki/index.php/Results#Using_k-means_and_trim ***** 11-24-2009 ver. 1.5.0.20 Opens .acquire files from Bruxton (TAC). Thanks to Dan Brown and Bruxton. ***** 11-17-2009 ver. 1.5.0.18 Model Report: Mean first passage between each pair of states bug fix: curve fitting: custom equation: editing parameters ***** 10-06-2009 ver. 1.5.0.17 to open files, drag-and-drop them from windows into qub ***** 09-02-2009 ver. 1.5.0.16 bug fix: File -> "Save idealized with event stats": works with large files ***** 08-20-2009 ver. 1.5.0.15 Optional qubcuda.dll (separate download) can greatly speed up Mac LL of large models ***** 07-30-2009 ver. 1.5.0.14 Double-clickable model files MSL bug fix and source commentary ***** 07-21-2009 ver. 1.5.0.12 New algorithm: Maximum Subinterval Likelihood (MSL): - like MIL, optimizes rate constants given idealized data - like Mac, a model variable (e.g. Voltage) can be driven by another A/D channel, also idealized -- assign Channel in data properties: data. ***** 07-13-2009 ver. 1.5.0.11 Much faster simulation with ramp protocols. note: if StimSim overflows (wraps around) on a voltage protocol, adjust Scaling in data properties, then re-StimSim. ***** 07-12-2009 ver. 1.5.0.10 bug fix: "Idealize" properties were initialized but the boxes weren't properly (un-)checked. ***** 07-06-2009 ver. 1.5.0.9 bug fix: ABF scaling (setting Offset to 0 to avoid baseline/offset conflict) ***** 06-29-2009 ver. 1.5.0.8 bug fix: Extract to DWT ***** 06-22-2009 ver. 1.5.0.7 bug fix: hang when reading some ABF files ***** 06-18-2009 ver. 1.5.0.6 Correctly read many more ABF/pClamp files. Please send us any ABF file which doesn't work. ***** 06-15-2009 ver. 1.5.0.5 bug fixes: Mac open conductance and "A Matrix" message ***** 06-12-2009 ver. 1.5.0.4 bug fix: IdlBase correctly handles multiple segments. ***** 06-11-2009 ver. 1.5.0.3 In MultiFit, after FitAll, subsequent KeepIt or FitAll replace the prior entry in the fits list clarification: when simulating with i=g*(V-Vrev) (model amps in pS), there's also a field to set baseline amp/std that's not V-sensitive bug fixes: Simulation and MacRates, if model amps in pS, expect voltage in mV Simulation, if model amps in pS, has correct noise profile Data (hi-res) curve fitting: removed "red scribble" at curve boundaries ***** 06-04-2009 ver. 1.5.0.2 bug fixes: Idl/Base amplitude levels resolved some 'Invalid A Matrix' messages ***** 06-01-2009 ver. 1.5.0.1 bug fixes: MIL floating point error window sizes when DPI != 96 stat: file list ***** 05-14-2009 ver. 1.5.0.0 More stable than 1.4.0.99, and recommended for everyone. ***** 05-12-2009 ver. 1.4.0.144 MIL parameters are internally scaled -- better when rates have different magnitude Fixed suggested A/D Scaling when not already known Fixed divide by zero in Data Overlap window. ***** 04-29-2009 ver. 1.4.0.143 Data tool-tip shows "Var / I": variance divided by mean, for selected data. ***** 04-24-2009 ver. 1.4.0.142 Bug fixes: Extract file list of selection lists Extract resample frequency textbox glitch Copy selection list items to clipboard MIL Word Report: kEq ***** 04-06-2009 ver. 1.4.0.141 New button: Edit Idl: apply dead time, edit classes, fill in blanks ***** 04-01-2009 ver. 1.4.0.140 Windows Vista compatibility: newer python/com installer right-clicking a model state brings up popup menu ***** 03-31-2009 ver. 1.4.0.139 Faster chopping Simulation bug fixes ***** 03-20-2009 ver. 1.4.0.138 Extract: File list -- to concatenate or average data files ***** 03-17-2009 ver. 1.4.0.137 Simulation bug fix ***** 03-13-2009 ver. 1.4.0.136 Bug fixes since 1.4.0.135: open data with multi a/d channels run QuB under WINE (linux) ***** 03-10-2009 ver. 1.4.0.135 "Global Fitting" window for dynamic clamp "Trigger Graph" window for dynamic clamp KPL scripting improvements and "Scriptlets" window "Chart" window Reorganized and expanded "Extract" options Resizable open and save dialogs Removed warning message from Results:Models browsing Data: Right-click to set the "Erase buffer", or to erase Shift and expand selections in a list Improved "overlap traces" display mode Stimulus processor bug fix ***** 02-25-2009 ver. 1.4.0.134 Fixed simulation with pS amplitudes, so both voltage and Vrev are given in volts, and pA = pS * (V - Vrev) ***** 02-19-2009 ver. 1.4.0.133 Right-click Results:Histograms to copy x or y coordinate to clipboard. ***** 02-06-2009 ver. 1.4.0.132 New 'Stimulus' processor -- improvement on IdlStim with visual preview and experimental tension modeling. ***** 02-03-2009 ver. 1.4.0.131 bug fix: Idealize with Half-amplitude (threshold): works with downward openings. bug fix: Results: Select: Make list(s) of selected: still works after renaming data. ***** 01-26-2009 ver. 1.4.0.130 Data properties: Data: Offset -- optional constant per-channel added to data points e.g. to interpret an a/d channel of zeros and ones as 1s and 5s instead, Offset = 1, Scaling = .25 ***** 01-15-2009 ver. 1.4.0.129 IdlStim: can filter stimulus with Spring-Dashpot model ***** 01-06-2009 ver. 1.4.0.128 Idealize: clearer presentation of "fix amp" and "fix std" options Dose-Response simulation: can "repeat each sweep" multiple times: - the model start state is reset only on the first rep - the peak is measured in each rep - the peaks are fit to an exponential ***** 12-18-2008 ver. 1.4.0.127 Dose-Response simulation: optionally runs two different stimulus variables. ***** 12-15-2008 ver. 1.4.0.126 Dose-Response simulation: optionally saves measurement results as text - suitable for fitting with QFC ***** 12-10-2008 ver. 1.4.0.125 selection list labels, when closely spaced, are now vertically staggered so you can read. If they still overlap, try Ctrl-+, to make the trace taller, and Shift-+, to make it wider. bug fix: Idealization with all these options chosen: 1) Half-Amp 2) Half-Amp stats 3) Apply dead time to idealization ***** 12-04-2008 ver. 1.4.0.124 installs and runs under Wine (Windows compatibility layer) correctly displays negative scaling factors [V / unit] MIL solves more Tcrit bug fix: Idealization (SKM) re-estimates all amps by default ***** 11-25-2008 ver. 1.4.0.123 new simulation wizard: Dose-Response (DRsp) runs a pulse train and plots the peaks bug fix: rename a selection list, and its name changes in the Data Overlap window. bug fix: MIL handles empty segments more gracefully ***** 10-31-2008 ver. 1.4.0.122 Fixed: in Extract: works with selection lists whose entries cross segment bounds. ***** 09-17-2008 ver. 1.4.0.121 Fixed: in Data: Copy Image: scale bars are no longer foreshortened by the top margin. ***** 09-12-2008 ver. 1.4.0.120 Fixed: MIL with smooth binning of very long events: not so slow any more. ***** 08-29-2008 ver. 1.4.0.119 Model Report works with large #channels (uses channel count: 1 instead of hanging) Model Properties: kinetic constraints: window resizes correctly ***** 08-27-2008 ver. 1.4.0.118 Warning when a stimulus channel is out of range Installs into the start menu for all users Model Report allows models using either pA or pS Chop and ChopIdl show the Data Source for clarity ***** 08-25-2008 ver. 1.4.0.117 Bug fix: opening .dwt files as data ***** 08-11-2008 ver. 1.4.0.116 Faster simulation of models with very many states (> 64). ***** 08-06-2008 ver. 1.4.0.115 Little fixes to Simulation, SimStim, Mac, and Survey ***** 08-04-2008 ver. 1.4.0.114 Survey: added Vrev, Cond and CondStd Fixed problems with model rate layout on saving Added model switch: pA (current) or pS (conductivity) for voltage-sensitive Mac Hid rarely used state properties ***** 07-28-2008 ver. 1.4.0.113 Fixed problems with model rate custom format strings (see 1.4.0.108) Restored scroll wheel behavior to pre-1.4.0.102 for "diagonal rates" style. ***** 07-24-2008 ver. 1.4.0.112 Improved reading of .prt files - can give one or more Acrit thresholds for closed/open and sublevels ***** 07-21-2008 ver. 1.4.0.111 Can open .prt (printout) files from David Colquhoun's EKDIST. Thanks to Steve Traynelis. ***** 07-10-2008 ver. 1.4.0.110 Fixed bug in Extract with join segments ***** 06-27-2008 ver. 1.4.0.109 Simulate: improved Gating Current (tiny--to see it, set gating units to 1e18 or more) ***** 06-26-2008 ver. 1.4.0.108 Mac: "min abs stimulus" can keep 0.0 from the auto-2level-idl-stim IdlStim can handle 1e-5 notation Simulate: optional Gating Current (tiny--to see it, set gating units to 1e21 or more) Model display options: Pre format and Exp format, "printf" style; blank for old behavior. ***** 06-11-2008 ver. 1.4.0.107 New button: "Survey" the results of MIL/Mac when stepping a parameter through a range ***** 05-24-2008 ver. 1.4.0.106 Optionally copy or print Model in color or black & white Data browsing stability ***** 05-23-2008 ver. 1.4.0.105 Results:Summary (and Excel Reports) include data source description "Edit constant baseline" in data right-click menu "bline apply" applies the baseline nodes to themselves IdlStim remembers the last A/D channel chosen Bug fixes. ***** 05-12-2008 ver. 1.4.0.104 Stability (Stbl) button generates moving-window event stats. Bug fixes. ***** 05-06-2008 ver. 1.4.0.103 Bug fix: edit data Scaling (bugged since 1.4.0.90) ***** 05-05-2008 ver. 1.4.0.102 Mac: choice of ExcelRow or ExcelCol cumulative reports ***** 05-02-2008 ver. 1.4.0.101 Testing Lorin's changes. ***** 04-02-2008 ver. 1.4.0.99 File -> Idealized Data -> Save idealized with event stats - like "save idealized data" - plus columns for mean amplitude and std.dev File -> Idealized Data -> Save idealized data as ASCII - added a column for source-event-duration ***** 01-23-2008 ver. 1.4.0.98 Results:Select:K-Means-Random: increased maximum K to 100. ***** 01-21-2008 ver. 1.4.0.97 Fixed bug in Mac LL calculation ***** 01-12-2008 ver. 1.4.0.96 Fixed bug in Results:Select (introduced in 1.4.0.95) ***** 01-02-2008 ver. 1.4.0.95 Model-merge - can specify the count of each source model Results - "Select" tool-tip includes the median To close several model or data files: ctrl-click in the list to hightlight all several, then right-click:close Idl/Stat/etc include "median_lifetime i" ("lifetime i" is the mean duration of events in class i) ***** 12-28-2007 ver. 1.4.0.94 MultiFit results include var_start for each var (a/d channel) ***** 12-19-2007 ver. 1.4.0.93 Fixed crash in Idealize etc caused by filtering too-small segments ***** 11-15-2007 ver. 1.4.0.92 Data print/copy now honors the "draw idealized in colors" setting when not in black and white New keyboard shortcut in data window: 'J'oin idealization: - extends the first event across the entire selection Better errors, warnings and double-checks in MIP, MIL and Model Merge Changed frequency histograms to log10(pA/Hz) and fixed a bug In Model:Kinetic Constraints: Changed "Del dependent" to "Delete extras" for clarity Added "Mac" as an option in Prtb (perturb rates) ***** 10-18-2007 ver. 1.4.0.91 Model editor forces k1 to 0 when Q is not checked, to avoid surprises. MIP, Mac, and IdlBase warn when the rates are too small or big IdlStim option: "Min duration" to avoid transient levels Histogram window: duration histogram generates error messages ***** 10-17-2007 ver. 1.4.0.90 "Prtb" (perturb rates) has a Word report ("Top 10") Idealization is cleared when you change scaling new "Delete entire idealization" in data r-click menu IdlStim latency in milliseconds Fixed MIL bug introduced in 1.4.0.89 ***** 10-15-2007 ver. 1.4.0.89 MIL etc handle k0=0 more gracefully (treat it as 1e-10 rather than freaking out) "Prtb" (perturb rates) has new option to use MIP instead of MIL ***** 10-12-2007 ver. 1.4.0.88 "Extract" now includes the idealization. ***** 10-05-2007 ver. 1.4.0.87 Option in ChopIdl to balance the event counts: dropping the last event if it matches the first, to get an equal number of open and closed events for stats. ***** 10-02-2007 ver. 1.4.0.86 Double-click the word "calculator" to clear the calc field ***** 09-27-2007 ver. 1.4.0.85 Idl/Base results moved to Data results list Default save data as QDF ***** 09-26-2007 ver. 1.4.0.84 Model search (msearch) new option: keep only top 20 models - fixes out-of-memory error in large searches. ***** 09-25-2007 ver. 1.4.0.83 MIP results and word report ***** 09-19-2007 ver. 1.4.0.82 Warnings about constraints: duplicate (linearly dependent) impossible (unbalanced) stimulus dependence in balanced loops "Idl show" re-checked when you idealize anything Net rate constants in rate tool-tip, as "k" ( k=k0*P*e^(k1*Q) ) ***** 09-18-2007 ver. 1.4.0.81 minor fixes ***** 09-14-2007 ver. 1.4.0.80 Hi-res data tool-tip is back ***** 09-12-2007 ver. 1.4.0.79 new "Model Search Top 10" report in MSearch results copy of MSearch results saved next to data file; reopen with File->Open Results ***** 09-06-2007 ver. 1.4.0.78 Bug fix: sampling rate is no longer corrupted in data:new,open,properties dialogs. ***** 09-04-2007 ver. 1.4.0.77 Mac: no need to idealize binary stimuli. If automatic 2-state Baum-Welch looks bad, use "IdlStim" under "Preprocessing" ***** 08-30-2007 ver. 1.4.0.76 Windows Vista compatibility: now you can close QuB. Please let us know how it works. ***** 08-29-2007 ver. 1.4.0.75 Bug fix: crash in AMP with more than two classes in the model ***** 08-20-2007 ver. 1.4.0.74 "Data overlap" window: added "sum" display option "Copy image" in the results histograms: fixed display corruption ***** 08-08-2007 ver. 1.4.0.73 Idealization (SKM) results include Filter freq in the summary tab. Warning when you delete or insert data: it could invalidate selection lists (they are kept as datapoint indices). Modeling:Clsf (Classify) shows a progress window. ***** 07-18-2007 ver. 1.4.0.72 ChopIdl results include (per output burst): - src_segment -- index of source segment - src_time -- start offset (ms) in source segment - src_ord -- ordinal number of burst in source segment, starting with 1 ChopIdl generates an additional "ChopIdl Src" result, with: - burst_count -- number of output bursts from input segment - burst_dur_mean -- mean duration (ms) of bursts this input segment - burst_dur i -- duration of i'th burst this input segment - burst_start i -- start offset (ms) of i'th burst this segment ChopIdl now drops the final event in a burst if it is the same as the first event. Thus, statistics such as Popen are calculated from an even number of open and closed events. ***** 06-21-2007 ver. 1.4.0.71 "LL" is back in the Idealize/SKM results (under Segments) ***** 06-18-2007 ver. 1.4.0.70 fixed: "Access violation" when editing parameters in Curve Fitting ***** 03-09-2007 ver. 1.4.0.69 fixed: Data low-res tool-tip was reporting wrong amplitudes ***** 03-08-2007 ver. 1.4.0.68 StimSim - Kept the Parameters up-to-date, activated the Help button Idealize, Idl/Base, Stat, ChopIdl: - added an option to apply the dead time (upper-right) to the statistics (recommended) - Idealize can additionaly (optionally) apply it to the idealization itself (not recommended) ***** 03-06-2007 ver. 1.4.0.67 Experimental multi-thread Mac StimSim - Simplified simulation with a stimulus waveform (either Acq.Protocol or Data file) to get started, create or open an Acquisition Protocol, then click "Stim Sim" "Presets" menu remembers and reopens protocols Mac and Exp remember all their properties ***** 03-02-2007 ver. 1.4.0.66 Can read ASCII text files in which the segments (traces) are in a column format Data files now remember the y range New display mode, experimental: "draw list only, overlapped" in Data Properties: Display if a selection is added to the list, all the selections are added in a new list Improved correction for leak and capacitance artifacts (LKC): now follows changes in voltage not limited to only one voltage step as before, one must do the LKC fit first with the Exp fitting feature, then Extract Added option for how many recent files are remembered (Options -> Remember MRU Files) Added "Scriptlets" window for use by running KPL scripts Copy/Paste as segment: preserves segmentation Mac: added ExcludeList option fit file list, each with own Fit/Exclude list Model: delete rate deletes any constraints involving that rate Constraints can be edited in-place, with a new Update button ***** 02-27-2007 ver. 1.4.0.65 Curve fitting: replaced "exponential (log10 x)" with "exponential (log x bin)" now you can enter the LogBase when fitting log-binned histograms you can type "e" as the LogBase for natural logarithms Data window: press 'M' to reMove selected baseline nodes ***** 02-26-2007 ver. 1.4.0.64 Solved crash when a selection list has items of different lengths ***** 02-23-2007 ver. 1.4.0.63 "Save data as...": progress bar and stop button MultiFit improvements ***** 02-21-2007 ver. 1.4.0.62 New MultiFit (MFit) button to fit each selection in a list ***** 02-09-2007 ver. 1.4.0.61 no more "crawling through" a new selection list ***** 01-11-2007 ver. 1.4.0.60 fixed simulation movies experimental data loading rewrite ***** 11-21-2006 ver. 1.4.0.59 fix for disappearing "session" info ***** 11-08-2006 ver. 1.4.0.58 new curve fitting: weighted improved display Results: Select: K-means improved ***** 10-27-2006 ver. 1.4.0.57 new curve fitting: shows components separately fits constant or linear baseline ***** 10-26-2006 ver. 1.4.0.56 new interface for curve fitting data, histograms, results fixed confusion between sampling kHz and msec in "New data" dialog ***** 10-11-2006 ver. 1.4.0.55 the model rate editor window is resizable ***** 09-06-2006 ver. 1.4.0.54 little bug fixes: Model Merge and IV ***** 08-31-2006 ver. 1.4.0.53 bug fix: showing Model Merge dialog bug fix: quitting QuB, then "cancel" on unsaved data/model ***** 08-25-2006 ver. 1.4.0.52 Enhanced wikiality: [Help] buttons load relevant editable pages from the web site We think data acquisition is stable. Prove us wrong. ***** 08-21-2006 ver. 1.4.0.51 Extract: save ascii with "segments in columns" ***** 08-17-2006 ver. 1.4.0.50 fixed obscure "node data" bug ***** 08-16-2006 ver. 1.4.0.49 instrumented to track down "node data" bug: if it happens to you, please send us the file C:\program files\qub\profile_log.txt ***** 08-14-2006 ver. 1.4.0.48 experimental data acquisition fix part 2: please test and report all crashes fixed bug editing filter and resampling frequencies in dialogs ***** 08-12-2006 ver. 1.4.0.47 experimental data acquisition fix: please test and report all crashes fixed crash with extremely small filter frequencies corrected baseline/filter processing in data overlap window: append avg/fn to data ***** 07-24-2006 ver. 1.4.0.45 custom histogram copy/print ***** 06-05-2006 ver. 1.4.0.44 a/d scaling calculator ***** 05-15-2006 ver. 1.4.0.42 bug fix: chop: no longer "walks" along displaying each selection ***** 05-08-2006 ver. 1.4.0.41 bug fix: reading ascii (txt) data bigger than 1 MB bug fix: reading corrupt (truncated) QDF data ***** 04-18-2006 ver. 1.4.0.40 added "Rename Results" to the Results list popup menu fixed hang when extracting a list of data channels fixed artifacts when extracting data filtered and resampled ***** 04-17-2006 ver. 1.4.0.39 bug fix (since 1.4.0.34: "node data index out of loaded bounds" when running Mac) ***** 03-25-2006 ver. 1.4.0.38 bug fix (since 1.4.0.33: "node data index out of loaded bounds" when Idealizing) simulated data takes on the model's name simulation by default replaces the last data simulated with the same model "Live Sim" re-runs the last simulation whenever the model changes - (only if both model and data are in front) Extract to DWT: Join segments fills in the blanks with class 0 (closed) dwells "Classify" button lets you divide a list according to which model fits each selection best Model: Kinetic constraints: "Del dependent" button removes constraints that are implied by others in the list. ***** 01-25-2006 ver. 1.4.0.37 Corrects errors in installer ***** 12-27-2005 ver. 1.4.0.36 proper renumber model states more accurate fit curve for mac skip slow changes fixed results:select:color:make selection list fixed crash when simulating stimulated by file fixed idealization anomoly with repeated large simulations simulation stimulated by proto file makes better estimate of a/d scaling improved mil smooth binning using triangular estimate ***** 12-12-2005 ver. 1.4.0.35 fixed SKM: Drop first last / if class fixed MIL sqrt histograms: using sqrt(sum(components)) not sum(sqrt(components)) Model copy image font shrinks to fit bitmap width Chop - ability to grab start and length from current selection Data selection list items : Single click to select also highlights selection ***** 12-01-2005 ver. 1.4.0.34 better mil histogram smoothing (better approximation to erf) Removed stim latency and resolution from mac, mip properties - (to make it clear you should manually IdlStim if you want to reidealize) - and automatically pop up the IdlStim window if there's no stim idealization. Channelcount on front panel, changes to channelcount management. ChannelCount stays exclusively in the model unless you're global fitting with the file list. Better names for instantiated results models MIP and Mac use the visible idealization of stimulus channels (in Mac it's optional for testing) New idealizer: "IdlCond" under "Preprocessing": - intended for stimulus idealization - invoked automatically if you mip or mac and stimulus isn't idealized - idealizes to known levels and/or levels separated by a delta MIL "sqr root ordinate" option Restored the output sel.list from ChopIdl Made building large sel.lists faster model revert (right-click the model file list) model geometry fix (hopefully all the labels stay in the right place now) model prevents you from serious edits during macLiveFit ibjumps, ext-sub-list, chop, goto, dynclamp, time markers, copypaste-sellist, (and more?) mil pdf:added option: "Smooth binning"; adjusted bounds of first bin to be part of geometric sequence - fixed upside-down-component bug - assumes (based on simulated evidence) that events measured as k samples are distributed normally (k*dt, dt/2) between (k-1, k+1)*dt stability: mac live fit with stop tasks and close model segment labels in data auto-labeled sels in lists data fit curve obeys "idl show" and "above" checkboxes mac results and reports extract filtered is correct at buffer bounds made fewer and larger buttons the default; small changes to default layouts wordreport option: rate-table that (re)calculates rate with stimulus wordreport tweak: rate-table formats with '%.4g' new data and model: base name plus timedate stamp some results (skm, stat, msearch) auto-held FOO Results auto-sel-list is no more results:select: make lists for all colors; in Select button/menu data overlap detail view includes fit curves ***** 09-26-2005 ver. 1.4.0.33 - Model merge with P and Q ( ligand and voltage ) handling. ***** 09-09-2005 ver. 1.4.0.32 - Repairs related to crashes during model building. - Point mouse to a rate, then use +/- to change for those without a mouse wheel. ***** 08-15-2005 ver. 1.4.0.31 - Simulation stimulation file can be a protocol file (no need to generate intermediate data file) - Idealize with half amp : can idealize data with a multiple channel model - bug fix : Save data always enabled; formerly could not save data when only activity was analysis - Ability to automatically add loop balance constraints in model ( properties -> Kinetic Constraints ) - MSearch option added to balance all loops - Word Report : Segments table added which outputs the contents of the results->Segments table - Word Report : Color means -- the mean and sd of each color each variable, in table form. - Black & white mode available on Energy Plot ***** 08-03-2005 ver. 1.4.0.30 - Mouse scroll wheel in model preserves constraints - Extract for multiple selected channels fixed. - Extract for large selections fixed. - Enforce constraints checkbox on rate - causes any edit of rate to enforce constraints - Large model hint shows only when the mouse is in the top 25 pixels of the model window - Note : Synch I/O, when stopped - may appear to hang while applying segment breaks; give it time it will complete. ***** 07-12-2005 ver. 1.4.0.29 - Mouse scroll wheel rate editor - repair for stimulated simulations and experimental conditions list - repairs / new options for MAC - various minor changes ... ***** 06-22-2005 ver. 1.4.0.28 - Warnings when attempting baseline adjust on channels > 0 - Random order option on protocol generation - Restored Model report - Multiple scale indicators on data copy image - Autocorrelation script for idealized data ***** 06-14-2005 ver. 1.4.0.27 - Bug on Simulation stimulated from file fixed - Changes to Energy Landscape plot - More improvements to National Instruments mx driver Including Synchronized I/O segmenting ***** 05-31-2005 ver. 1.4.0.26 - Correct reading of units for ABF float data - Ability to change state indices in models ( Rt-click, Set Index ) ***** 05-24-2005 ver. 1.4.0.25 - Fixed length synchronized I/O bug fixed. ***** 05-23-2005 ver. 1.4.0.24 - National Instruments NIDAQ-mx support - Synchronized Input / Output ( for NI-MX driver only ) - MAC algorithm improvements - File Templates - Numerous small changes... ***** 03-01-2005 ver. 1.4.0.23 - Fix : Digital output sending brief incorrect data during output. ***** 2-10-2005 ver. 1.4.0.22 - Fix : Digital output sporadically causing lockup. - Fix : Digital output sporadically causing lockup. ***** 2-3-2005 ver. 1.4.0.21 - Inclusion of digital output (from ver .19) was in full version only, now in upgrade too. ***** 1-19-2005 ver. 1.4.0.20 - Idealized - export to clipboard for cut and paste with excel, igor, ... ***** 1-17-2005 ver. 1.4.0.19 - Digital outputs in protocols. ***** 12-30-2004 ver. 1.4.0.18 - Repairs, speed improvements. ***** 12-01-2004 ver. 1.4.0.17 - Repairs some MIL / SKM bugs introduced in version 15 - New rate edit mode is completed ***** 11-18-2004 ver. 1.4.0.16 - Rates edit screen allows selection of rate. Click on [*] button to try it. - Numerous internal bug / efficiency / memory leak repairs. Be alert for new side effect problems. - Ability when opening Ascii or ABF data to have QUB 'pre-scan' part of the file to estimate scaling. - Tooltips added for active channel, import scaling, several other items. - Changes to output text for MAC algorithm in 'quiet' mode. - Movie mode in energy model for viewing effects of stimulus dependance. ***** 10-06-2004 ver. 1.4.0.15 - added reading of IGOR .ibw files with 4 byte floating point data. (WWW.WaveMetrics.Com) ***** 09-16-2004 ver. 1.4.0.14 - Fixed ability to view output protocols ***** 09-16-2004 ver. 1.4.0.13 - Internal maintenance ***** 09-07-2004 ver. 1.4.0.12 - Internal maintenance ***** 08-18-2004 ver. 1.4.0.11 - Pseudo T(s) added to model pEq calculation - Repaired a bug which occurred with SKM on multiple segments, in versions 1.4.0.9 and 1.4.0.10 only. ***** 08-18-2004 ver. 1.4.0.10 - Ability to partially import Synaptosoft model files. See also WWW.Synaptosoft.com. ***** 08-11-2004 ver. 1.4.0.9 - Better compatibility with ICE idealized data files (.dwt) - MAC algorithm - remembers parameters, allows for 'latency' - 3d Energy model has been enhanced ***** 07-28-2004 ver. 1.4.0.8 - In some cases IDL was taking very long and not producing histograms. - Energy Model minor changes - max peak ln(k0) changed from 10.0 to 14.0; Toggle state cubes; Current state in model is zero state. - Kinetic Constraints : larger window, better error feedback, copy to controls on delete provides ability to edit via Delete+Add. ***** 07-23-2004 ver. 1.4.0.7 - Heka files were not correctly loading some parameters, particularly sampling rate. Corrected. ***** 07-12-2004 ver. 1.4.0.6 - Scrolling of reports window corrected - Right click on model - select Energy Model - displays a 3-d energy model. This will be enhanced next month. ***** 06-30-2004 ver. 1.4.0.5 - Internal maintenance ***** 06-17-2004 ver. 1.4.0.4 - New Task: MdlRpt (for real this time) ***** 06-16-2004 ver. 1.4.0.3 - New Task: MdlRpt - displays Q,A matrices, Peq, time constants, mean current, etc. ***** 06-04-2004 ver. 1.4.0.2 - Fixed bug in Simulate: Idealized with Output: State ***** 06-02-2004 ver. 1.4.0.1 - Made F1 more likely to find a help topic. ***** 05-27-2004 ver. 1.4.0.0 - Official QUB 2004 Release ***** 05-20-2004 ver. 1.2.0.86 - Tcrit lines on MIL histograms are visible again. ***** 05-19-2004 ver. 1.2.0.85 - Fit curves from Mac and Exp are visible again. ***** 05-18-2004 ver. 1.2.0.84 - Calibration assistant (options menu) for data acquisition setup. - Upgraded Python/COM extensions. ***** 05-10-2004 ver. 1.2.0.83 - Bug fixes ***** 04-30-2004 ver. 1.2.0.82 - MIL displays the mean time constant for each class in the Report window - SKM responds more quickly to "Stop Tasks" ***** 04-29-2004 ver. 1.2.0.81 - fixed data curve fitting - fit curves from Mac and Exp are temporarily invisible ***** 04-28-2004 ver. 1.2.0.80 - can open .evl files (Axon event lists) - keeps correct input channel calibration (scaling) for different gain settings. ***** 04-22-2004 ver. 1.2.0.78 - Interface menu -> Additional layouts ***** 04-21-2004 ver. 1.2.0.77 - new button: Exponential fitting of macroscopic currents ***** 04-14-2004 ver. 1.2.0.75 - Action buttons are accessible from the menu. - Redesigned Acquisition Hardware setup - National Instruments cards setup recognizes many card types and sets default parameters. ***** 04-12-2004 ver. 1.2.0.73 - "Fix all rates" button in Model properties ***** 04-07-2004 ver. 1.2.0.72 - Extract ASCII: bug fix ***** 04-06-2004 ver. 1.2.0.71 - "Extract": can save an arbitrary subset of A/D channels. ***** 04-06-2004 ver. 1.2.0.70 - "Extract": bug fixes. ***** 04-01-2004 ver. 1.2.0.69 - Idealize/SKM bug fix ***** 03-23-2004 ver. 1.2.0.68 - MIL works with multiple channels and subconductance levels. - new model constraint: "Loop imbalance" preserves the ratio of the product of forward rates to the product of backward rates in a loop. - bug fixes ***** 03-11-2004 ver. 1.2.0.66 - Included sample acq protocols in QUB\Sample Protocols (to open, click the second icon in the Acq Protocol window) - Fixed a protocol bug introduced in 1.2.0.65 ***** 03-10-2004 ver. 1.2.0.65 - Reads more varieties of DWT file - Acquisition protocol improvements: - custom shapes -- f(x, i) for x in [a, b] - geometric progressions (using e.g. "Amp factor") - zoom n' scroll - see the n'th repetition by setting "Repeat stage" to n ***** 03-02-2004 ver. 1.2.0.62 - Results: Reports: added "Select: Charts" option ***** 02-25-2004 ver. 1.2.0.61 - Model: Copy Image: "Named states" option - labels states like C1 <---------> O1 <------> O2 - intended to create a meaningful black & white image -- if you have suggestions, email contact @ www.qub.buffalo.edu - Results: - only the most recent results are kept - to keep a result indefinitely, right-click its entry in the Data window and choose "Hold Results" - Select: - to make a selection list of e.g. all the red segments, right-click the color and choose "New Selection List..." - Reports: - click the button to build your own custom report - you can program new options for the reports in Python (for examples, look at the second half of WordReport.py) - to share a report format with friends, copy the file from \Program Files\QUB\Reports\ ***** 02-20-2004 ver. 1.2.0.60 - Acquisition Protocol aesthetics ***** 02-12-2004 ver. 1.2.0.58 - stability fixes ***** 02-10-2004 ver. 1.2.0.57 - "Extract" can write .qdf files - New "Find Notes" searches for files by the contents of their "Data Notes" - To "grab all amps" from the selected data, right-click the background of the model - "Chop Data" and "Chop Idl" (EvtP) act on the "Data Source" - "Chop Data" can discard the last segment if it's too short ***** 01-29-2004 ver. 1.2.0.56 - Acquisition Protocol: - shift-click to select multiple shapes - improved param editor ***** 01-27-2004 ver. 1.2.0.55 - To show/hide buttons on the right, right-click a category label (e.g. Modeling). - In the Data Overlap window, you can zoom out by right-clicking the detail view. - Curve fitting has more built-in curves. - New Mac(roscopic) layout button. - The Data Comment window has been renamed to "Data Notes". - Simulate now works as advertised. Specific fixes involve stimulation files, channel count, and parameters. - The Data Overlap and Acquisition Protocol windows keep their proportions when resized. ***** 12-12-2003 ver. 1.2.0.50 - New "Acquisition Protocol" window lets you create stimulation protocols graphically. - To create a protocol: - press the top-left button to start a new protocol file - drag shapes from the row of buttons into the central area - click a shape and edit its parameters on the left (press enter to confirm each value) - a "Stage" can be repeated. "delta_" parameters tell how much to increase a value each rep. - To use it: - in the "Acquisition Output" window, choose Data: New - press the top-right button in the protocol window to generate sampled data for output. Do this again after changing the protocol. - press play in the Acq.Output window to send the samples to the equipment. - Simulation of stimulated data no longer requires a script. You can specify a data file (such as the output of an acquisition protocol) as the stimulus. - Use the "Parameters" tab to match model parameters (e.g. "Ligand") with A/D channels of the stimulus -- channel 1 is the first stimulus channel. - Make sure the output file has one more A/D channel than the stimulus. - New "Data Overlap" window shows overlaid and averaged traces. - You'll need a selection list of similar traces - if each trace is in a separate data segment, press "AllS" - if the traces are regularly spaced, try "Chop" - or build a list by hand with "AddL" and "AddS" (the 's' key == "AddS") - Choose your list in the Overlap window - Drag across the display to zoom in - Click a trace to highlight it in the list and vice-versa - Check/uncheck traces to include them in the zoomed display - Right-click the zoomed display to show an average or custom function - Right-click an average or function display to curve-fit or append it to the data file - Data Source: List refers to only the checked traces in the zoomed area. - New "IV" button plots the average level of each A/D channel vs. the active one (in "Results") - New "Stat" button generates the same info as idealizing (amp, occupancy, lifetime, ...), but for data that's already idealized. - Mac improvements ***** 10-16-2003 ver. 1.2.0.35 - Bug fix: you can re-start acquisition with NI boards. ***** 10-15-2003 ver. 1.2.0.34 - Each channel in a data file has a Name, Units and Scaling. - For data stored as integers: x (V) = intval / ADScaling [ADScaling = 2^bits / range (V)] y (Units) = x (V) / Scaling - Edit them in Data Properties:Data. - Model amps are in Units. - "Copy Image" works with the high-res data display. ***** 10-07-2003 ver. 1.2.0.33 - Bug fix: data points that are out-of-range no longer overlap with neighboring traces. ***** 10-04-2003 ver. 1.2.0.32 - Data Acquisition re-organization, part 1: - First setup hardware: Options menu -> Acq. Hardware - Then define and name channels: Options -> Acq. Channels - "Single-shot controls": View menu -> Acq. SShot: - send/recv a single value - first click Properties, choose which channels to monitor - click the color strip on the right to update. - Recording controls and scope: View menu -> Acq. Input: - first click Channels, choose which channels to record - click or r-click "Data: " to assign a file to record into - Output (protocol) controls: View menu -> Acq. Output: - first click Channels, choose which channels to write - click or r-click "Data: " to assign a file to play out ***** 09-24-2003 ver. 1.2.0.28 - Cosmetic stuff: - Results are listed (to the right of data) by name and data source - To clear out a bunch of old Results, shift- or control-click to multi-select, then right-click and choose Delete. Sorry I can't get the delete key working in the Results list. - Right-click a point in Results:Select to see its "segment number" ***** 09-24-2003 ver. 1.2.0.27 - Added "List show" checkbox at upper right. - In data properties, "Set as default" includes the dead time. ***** 09-23-2003 ver. 1.2.0.25 - Upgraded to Python 2.3 ***** 09-23-2003 ver. 1.2.0.24 - bug fixes: - better handling of corrupt session files (.qsf) - MIL works with .qmf model files - 'ChannelCount' in the model and the file list reflect each other. ***** 09-22-2003 ver. 1.2.0.23 - Data in the current selection list is highlighted. (to turn off the highlight or customize the color: right-click the data, choose Properties) - Each selection in a list can have a label and a comment: - right click a selection to edit its label - right-click the AddS (add selection) button to add a labeled selection. ***** 09-16-2003 ver. 1.2.0.22 - Correctly opens ABF floating point data. ***** 09-11-2003 ver. 1.2.0.20 - "Perturb MIL" ("Prtb") alters each rate in turn and tests the likelihood. ***** 08-26-2003 ver. 1.2.0.19 - Converted MIP, Mac, Stair to the new file list. - Put dead time (tdead) at upper-right. ***** 08-25-2003 ver. 1.2.0.18 - Converted AMP and MSearch to the new file list. - Fixed bugs from 1.2.0.17 with AMP and WordReport. ***** 08-25-2003 ver. 1.2.0.17 - MIL and MPL have a new file list. Both can now fit globally (across files) using models with one or more concentration params. ***** 08-04-2003 ver. 1.2.0.16 - Bug Fixes: - portions of the idealization were disappearing. ***** 08-04-2003 ver. 1.2.0.15 - Bug Fixes: - "Add a baseline node" works like you'd expect. - MIL/MIP/Drift works with data source: "list". - Session info (idealization, baseline, results) is preserved if you edit data but don't save changes. ***** 08-03-2003 ver. 1.2.0.14 - Features: - When you idealize a selection or list, it no longer clears the idealization in the rest of the file. - To erase some dwells by joining them with the previous dwell, select them, right-click, and choose "Join Idl". - Bug Fixes: - When you delete or insert data, the idealization stays correct. - You can edit a .qdf file after re-opening it. ***** 07-29-2003 ver. 1.2.0.13 - Saved baseline nodes include the sampling rate, so you can - extract the whole file, resampled down - correct the baseline of the resampled file - save the baseline nodes - go to the original data and load the baseline nodes ***** 07-11-2003 ver. 1.2.0.12 - Baseline nodes are not cleared when you delete or otherwise edit the data. ***** 06-27-2003 ver. 1.2.0.11 - Sorted out baseline correction. In the data's right-click menu: - "Set baseline" offsets the whole file by the average selected current. This offset occurs whether or not "Bline apply" is checked. - "Set baseline" also defines the "Baseline selection" (list), which is used to "Erase" a selection of data. - "Add a baseline node" works as before -- to correct simple drift, add a node at the beginning, a node at the end, and anywhere the closed current goes too far from the zero line. - "Auto" display scaling now uses the processed data (baseline, filter). ***** 06-24-2003 ver. 1.2.0.10 - Fixed "Unknown error" in MIL after saving the model. - Idl/Base histograms use baseline-corrected data - Model parameters (Ligand, Voltage) are set in the data file - to change a parameter: right-click the data, choose Properties, then Info. ***** 06-16-2003 ver. 1.2.0.9 - Model Merge: UnMerge preserves class colors - Idealization (SKM) bug fix ***** 06-16-2003 ver. 1.2.0.8 - Integrated model merge (right-click a merged model to un-merge) - MIL bug fix ***** 06-12-2003 ver. 1.2.0.6 - Results : Summary tab shows numbers like overall likelihood and amplitudes. ***** 06-11-2003 ver. 1.2.0.5 - Data->WAV (dWAV) converts sampled data to sound. - Fixed a bug in MPL idealization display. ***** 06-09-2003 ver. 1.2.0.4 - The "List Stim" button makes a selection list of the times when a stimulus channel is past a threshold. - Results are listed with their data file, above selection lists. - they will persist across QUB sessions. ***** 05-21-2003 ver. 1.2.0.2 - "Set baseline" warns you when it will clear existing baseline nodes. ***** 05-09-2003 ver. 1.2.0.1 - Extract: when adding or averaging selections of different length, the output is as long as the shortest selection. - MIL: displays "Last LL" in the text output - Model window: removed right-click:delete model ***** 05-05-2003 ver. 1.2.0.0 - Can extract "Select" histograms to Excel. - Displays newly opened files the same as the last opened file. ***** 05-02-2003 ver. 1.1.0.2776 - "Compute Stats" is optional in Idealize:HalfAmp and ChopIdl. Stats include amp, lifetime, and histograms. They can consume significant time and memory. ***** 05-01-2003 ver. 1.1.0.2775 - A bunch of bug fixes. ***** 04-30-2003 ver. 1.1.0.2774 - Correctly reads Pulse, ABF floating point, ASCII, and .dat files. ***** 04-28-2003 ver. 1.1.0.2773 - Fixed bug where all amps in model became 0.0. ***** 04-24-2003 ver. 1.1.0.2772 - Fixed crash in MIP, Mac, and Drift when clicking "Cancel" - Fixed crash in MPL - Data tool-tip updates as you move the mouse - MIL histogram pdf components can be toggled on/off - Right-hand buttons default to small, all visible in new installations. - Fixed "number 0" bug in qubtree.dll (It made all "Select" charts have log y axis.) ***** 04-23-2003 ver. 1.1.0.2771 - "Idealize" and the other buttons work on the active channel ("Ch" on the top bar). ***** 04-21-2003 ver. 1.1.0.2769 - MIL Word report includes a picture of the data. ***** 04-18-2003 ver. 1.1.0.2768 - Keyboard navigation of selection lists: arrow keys and - Results:Histograms have a tool-tip with the mouse coordinates - MIL: - Tcrit is shown as a red line in Results:Histograms - "Show Iterations" checkbox affects text output and histograms ***** 04-17-2003 ver. 1.1.0.2767 - Can read .abf floating point and HEKA .pul files (experimental) - The model can have any number of classes - A "session file" (.qsf) automatically preserves idealization, baseline, selection lists, and display settings - Spectrum histograms work more reliably - Idl/Base can re-estimate rates (option "A -> Q") - MIL: - set dead time in milliseconds or as a multiple of the sample rate - file lists can be saved and re-loaded - histograms have a line for each exponential pdf component - (now shown as [count / total] instead of its square root) - MS Word report includes pictures of the final model and the data ***** 04-07-2003 ver. 1.1.0.2766 - Improved baseline correction in Idl/Base. ***** 03-13-2003 ver. 1.1.0.2765 - "Chop Idl" (the event processor) now handles the end of a segment correctly. - "Idealize" can apply a dead time to its output. - The result of "Idealize" no longer includes the "a matrix". - Point the mouse at a rate in the model to see delta G in kT units. - Two new buttons, "MakeStar" and "MakeChain", build a model by adding states until the MIL likelihood stops increasing. - The source code is in QUB/PythonScripts/ModelBuilder.py. It's an example of how to add buttons and work with the model using Python. - QUB/Python API documentation is available at http://www.qub.buffalo.edu/qubdoc/ ***** 02-22-2003 ver. 1.1.0.2764 - Histograms for idealization and optimization have been moved to the Results window. - Fixed an off-by-one error in the calculation of amplitude histograms. (most noticeable with short events) ***** 02-19-2003 ver. 1.1.0.2763 - Added an option for old-style small buttons in the Interface menu. - Improved stability. ***** 02-18-2003 ver. 1.1.0.2762 - Fixed an intermittent crash after idealization, MIL, model search, and simulation. ***** 02-14-2003 ver. 1.1.0.2761 - Now you can customize the buttons on the right. - you can have different buttons visible in each layout. - from the "Interface" menu, choose "Edit Buttons". - "EvtP" is now known as "Chop Idl" ***** 02-10-2003 ver. 1.1.0.2760 - Fixed multi-part stimulation bug (simulation and data acquisition). ***** 02-07-2003 ver. 1.1.0.2759 - Fixed crash when idealizing huge files with SKM. - Improved aesthetics of Results:Select:AsHistogram. - Save and reload baseline nodes (File menu) as a .qbn file. - Data histograms (Amp, Spectrum, Duration) operate on the Sel/List/File as chosen on the right. - If "Sel" is chosen, they use the entire low-res selection. (They used to use just the visible part.) ***** 01-29-2003 ver. 1.1.0.2757 - Curve fitting is available for all histograms and result charts, and the high-res data window. - right-click the chart or data and choose "Curve Fitting" - you can add custom curves -- see QUB/PythonScripts/CurveFitting.py for details - to fit only the selected high-res data, right-click the selection and choose "Expand" so it fills the high-res window. ***** 12-30-2002 ver. 1.1.0.2753 - EvtP: the event processor can chop when an event is too short. ***** 12-24-2002 ver. 1.1.0.2752 - Various bug fixes ***** 11-26-2002 ver. 1.1.0.2747 - Python Scripts: - the threading module now works correctly - MIL - fixed pdf bug from 1.1.0.2745 - Results: Select - linear regression (right-click chart, "Curve Fitting") ***** 11-18-2002 ver. 1.1.0.2745 - For multichannel data (stimulated): you can hide one or more channels. - right-click the data, choose properties, choose a channel, then unselect "Visible." - EvtP (event list processor) - more options - generates a Result with the same stats as Idl - Results: Select: - plot one variable versus another by "Show Var..." -> "Custom..." - MIL: - made numerical over-/underflow less common - fixed bugs involving balanced loops and 2 or more channels - "Stop Tasks" also interrupts Python. ***** 11-07-2002 - MSearch (model topology search with MIL) is now included in QUB. - Make a model with the desired number of states for each color, then use the "msrch" button. - Results are in the Results window, under the "Models" tab. - To see a model, click its line. - To sort, click a column header. - To copy the table to the clipboard, click the upper-left corner then press Ctrl-C. - "normLL/seg scatter" is the std. error of (LL / num.dwell) across segments; "normLL/file scatter" is the same deal for files, if you're using the file list. ***** 10-10-2002 - To extract a Results chart to MS Excel, right-click it. - Conductance classes are consistently numbered 1 - 10. - In Idealize/SKM: the FixAmp[i] and FixStd[i] checkboxes refer to the same class #i as in the model. - When idealizing with more than one channel present, a multi-channel amplitude is fixed only if each contributing amp is fixed. - Improved stability and start-up time. ***** 10-02-2002 - You can right-click most histograms and "Extract to Excel" (if you have MS Excel 2000 or newer) ***** 09-30-2002 - It's now safe to run SKM on large data sets. ("Idl" button, "SKM" method) - Half-amplitude idealization generates histograms and statistics (the same ones as SKM). - To copy a picture of a Results chart to the clipboard, right-click it. ***** 09-23-2002 - See the model before and after optimizing, with the "Models" tab in Results. - MIL displays Tcrit for each consecutive pair of time constants: - Amp1 * exp( - Tcrit / Tau1 ) = Amp2 * (1 - exp( - Tcrit / Tau2 )) ***** 09-17-2002 - Select segments using inequalities with the "Criteria" tab in Results. - "Select" tab improvements include: - Show a variable with log y axis (right-click the chart) - "Trim" one color at a time (right-click the color) - Remove segments of a color, or extract them to a new Result - Position the pointer over a chart to see the mean and s.d. of each color - MIL Histogram extraction works correctly ***** 09-13-2002 - If QUB is showing a message about Microsoft Excel or Word, get some coffee and let it finish. It won't happen again. (It's setting up the MIL Word Report feature.) - Convert idealized data to sound: - Idealize some data then right-click "Sing" under Modeling. - Check the time scaling. (1 is actual speed, 2 takes twice as long) - When you click Run, you'll be prompted to set up a new data file. Choose: - File format: WAV - AD Channel Count: 1 - Sampling [kHz]: 11.05 - Then press OK. - You can hear the sound by pressing "Play" (lower right), or File -> "Save Data As" and listen with a standard music player. - Results View - To see Select charts as histograms, right-click them. ***** 08-29-2002 - Results View - Visualize and process the output of SKM, MIL, Amp, and MPL. - To see it: run MIL then click "Results" on the left. - MS Word reports have moved here. Click [Reports] for the menu. - The "Select" tab plots parameters by segment, and lets you choose subsets of data to work with: - Group segments together by coloring them (click or drag). - Segments of the chosen color are in a selection list named like "MIL Results" - To quickly separate segments into groups: - [Select] -> "None" - Color a sample point from each group (different colors). - [Select] -> "K Means" to classify the rest. - Can save and re-open result files. ***** 08-05-2002 - MIP: - A new single-channel optimizer. - Uses idealized data, like MIL, but considers points rather than intervals. - Like MPL, MIP does not need missed event correction. - Does not currently work with multiple channels. - Mac: - A new macroscopic current optimizer. - Estimates rate constants and number of channels. - MIP and Mac are under testing!!! - Both can fit multiple files across experimental conditions and accept constrained models. - For details, email Lorin Milescu (milescu@buffalo.edu) - MIL: - You can make a MS Word report after MIL finishes: Click Rpt (on the right under Mac) and choose "MIL Word Report".(see 8-29-02) To customize it, open QuB\PythonScripts\WordReport.py in a text editor. ***** 06-06-2002 - Simplified baseline correction. - Fixed MPL memory leak. ***** 05-30-2002 - "Stop Tasks" on the lower right will stop all long-running jobs (e.g. Idl). It's a shortcut for View -> Tasks -> Stop. - Amp can take its initial amplitude guesses from the high and low points of the data. - Right-click on Amp, then select "Auto-Init" - For inward current, choose "Open is Down" - If "Auto-Init" is not selected, the initial guesses will come from the model. You may need to "Grab" initial guesses from the data. ***** 05-21-2002 - New algorithm buttons: - Amp: Find optimal conductance levels using Baum-Welch. - MPL: Maximum Point Likelihood. Like MIL for nonidealized data. Slower than MIL. - You can choose a Data Source for Idl, Amp, MIL and MPL using the buttons just below them. - Python scripts - Various subtle bug fixes. ***** 04-16-2002 - MPL (maximum point likelihood) was added as an option in Mil. - MIL/MPL have new options: - Restarts: when it reaches max iterations, the optimizer is restarted. - Max Step: Amount by which the optimizer changes parameters each iteration. - Acquisition: - Pulses can be applied independently to each AD channel. - Digital pulse was added (same syntax as analog pulse). ***** 02-14-2002 - The new file format '.QMF' is available for model files. Under testing ! - The comma problem was fixed in the sense that QUB always uses the '.' as the decimal dot. ***** 01-07-2002 - The new Options menu lets you choose how qub starts up. ***** 12-21-2001 - Data Display: - the infamous DoubleClick bug was fixed !!! - traces can be inverted, for display-only (or printing) (Data.Properties.Display.Inverted) - the selection can be repositioned with the mouse (hold Ctrl, mouse left-click and hold down, and move the mouse to reposition the selection; Ctrl can be released, once the mechanism is initiated, but mouse must be pressed) - the selection in the higher resolution window can be expanded to the whole window, and un-expanded back to where it was (data.rightclick.expand/unexpand) - Selection Lists - DoubleClick on the data display selects the corresponding entry of the current sel list, if any - Delete on a selection list deletes the current entry if there is no selection, or all the selected entries, or, if Control is pressed, deletes all but the selected ones - Model Display - Edit operations are available now: Cut, Copy, Paste and Delete, operating on the whole model ***** 11-26-2001 - MIL: - can evaluate the current model for each segment separately. - generates time constants - find them in the Report window, the Word report, and by pointing at histograms - Models with negative exponential rate constants (b or k1) can be reopened. ***** 11-17-2001 - MIL: - when fitting just one file, set the voltage, concentration, and dead time in the MIL dialog box (right-click "Mil"). - MIL used to get these from the "MIL Input List" window; Now those defaults are only for files added to the list. - the Word report: - is now saved in the same folder as the (first) data file, as _rep_.doc - histograms are embedded Excel charts instead of pictures - LogLikelihood has the same sign as in classic MIL - more descriptive error messages ***** 11-12-2001 - You can fit histograms with up to 8 gaussian components: - right-click a histogram and choose "Curve Fitting" -> "Gaussian" - "Extract" saves the curves along with the histogram - only for the "Selection" histograms (amplitude, power spectrum, and duration) ***** 10-31-2001 - MIL new features: - can create a summary report in Microsoft Word; - use the option "Make Word Report" after right-clicking "MIL"; - you must have Word and Excel installed on your computer. - stops more quickly when you click "Stop" in the "Task View" window. ***** 10-24-2001 - All histograms can be copied to Clipboard as pictures (enhanced metafiles; RightClick, Copy); - Idl and Opt Histograms can be extracted (RightClick, Extract); - MIL bug fix (automatic loop balancing); - SKM improvement (baseline track); - ACQ (acquisition) bug fix in the Data Translation driver ***** 10-09-2001 - MIL new features: - can optimize segments separately; - use the option "Each segment" after right-clicking "MIL"; - the results are saved in a file called "MILreport.txt", located in the most recently-used folder. - SKM - corrects baseline better; - handles complex models correctly in large data and baseline modes. ***** 10-02-2001 - Baseline tracking in SKM was improved: - faster - parametrized by standard deviation of the baseline ***** 09-25-2001 - SKM new features: - can handle large data, with a cost in speed; check the option "Allow Large Data"; - can track (more or less) baseline wandering, again, with a (higher) cost in speed; check the option "Baseline Track" and set "Win". - MIL new features: - first of all, it works ! - can fit across files with different number of channels - shows histograms per file - The units in the Event duration histograms have changed. ***** 09-06-2001 - Histograms work now in QubMIL. ***** 08-23-2001 - MIL was added to QUB. Histograms don't work yet. ***** 06-22-2001 - Extract DWT is much faster for gigantic segments. ***** 06-21-2001 - When either data or model window becomes active, a focus rectangle is drawn. ***** 06-20-2001 Some features that were not described yet: - QUB registers *.dat, *.ldt, *.dwt and *.mdl extensions. If a file with that extension is double clicked in windows explorer or winfile then it is opened by qub. The command qub.exe -od datafile for data files and qub.exe -om modelfile for model files. - Chris played some tricks with the SKM idealizer and saved a lot of wasted memory, at least 30%, and more for more complex models. - A driver for Data Translation acquisition cards was added ("qubacqdx.dll"). ***** 06-20-2001 - Occasional error in Model equilibrium probability calculation fixed. ***** 06-15-2001 - Simultaneous drawing (technical stuff, please ignore) issue addresed, and hopefully fixed. Probable cause of some unexpected and unexplainable crashes. - Took the courageous decision of not locking data during acquisition. The effect is that any operation (except save and close) can now be performed on a recording file. In other words, on-the-fly analysis is now possible. Not tested yet, so do it at your own risk... ***** 06-14-2001 - Data recording was improved and some problems fixed. - Can extract segmented ASCII files. - Can read segmented ASCII files: segments must be separated by at least one empty line. ***** 06-14-2001 - This file started.