![]() |
EDIT menu
|
This describes the portion of the EDIT menu covering topics other than respirometry. See the Respirometry submenu for gas exchange calculations.
For BASELINE, INTEGRATE, DERIVATIVE, SMOOTHING, GAS EXCHANGE, and some other EDIT menu calculations, clicking the plot window is equivalent to clicking a highlighted button.
The small pop-up menu in the upper right corner of the plot area is one of several ways to select the 'active' channel from a list of available channels, which is updated whenever you alter the current channel labels (by loading a new file, editing a channel label, creating or removing a channel, etc.).Other methods for channel selection:
Use the number keys: Press the key that corresponds to the number of the channel you wish to use (press '1' for channel 1, etc.). Use the 0 (zero) key for channel 10. For channel numbers above 10 (to a maximum of 20), use a combination of the shift key and a number key (i.e., press shift 1 for channel 11, shift 2 for channel 12, etc.).
Use the up and down arrow keys: The 'up' arrow sets the active channel to the next higher channel number, and the 'down' arrow sets the active channel to the next lower channel number.
Note that this method will 'wrap' at channel 1 and at the highest-numbered channel. In other words, if the active channel is 1 and you press the 'down' arrow key, the active channel is set to the highest-numbered channel available. Similarly, if the active channel is equal to the highest-numbered channel and you press the 'up' arrow key, the active channel is set to 1.
In multiple-channel mode, use the buttons in the floating window: The pushbuttons on the floating 'toolbar' at right of the plot area indicate the channel in use (active channel) and let you select new active channels by button clicks. The other means of changing channels still work.
When performing a complex sequence of manipulations on data, remember that only one level of UNDO is available. This means you can only reverse the most recent change, and for some operations that affect all channels, NO UNDO is possible. |
By clicking one of the 'channel x' buttons, you can change to the selected channel. Comments can also be edited up to the maximum of 32K characters (240 characters if you want to save in Sable SSCF format).
Clicking on the 'comments' window in the main display takes you directly to the 'edit comments' window.
![]()
1. Automatic -- The program selects blocks at the start and/or end of the data and uses them in a regression to compute the baseline. You can change the number of cases in these blocks by editing the value in the box (the default is 15). Use the automatic method only if the beginning and end of the file contain nothing but true baseline values. If this condition is not met, use another method (such as #2 or #3, below).
2. Automatic (periodic) -- (shown at right) The program uses one of two methods to identify baselines: (a) you use the cursor to select a starting point and time interval between reference readings, and the program uses them to 'step' from reference to reference, or (b) the program uses markers in each reference as a cue for baseline setting (in my experience, this works considerably better than the first method). You can chose which marker is the reference indicator, and whether the reference adjustment should begin on either side of that marker. Only the latter method (using markers) can be scripted.
3. Use the mouse to select a sequential series of baseline blocks (multiple points), starting from the 'left' end of the file and moving to the 'right' end. This method corrects the baseline in a series of segments (up to 300 in a given file; you can repeat the process if you need to handle more than that). Alternately, you can click on single points (instead of selecting blocks).
4. Use the mouse to select beginning and ending blocks, which LabAnalyst then uses in a linear regression to compute the baseline. Alternately, you can click the single point button to pick baseline points (instead of selecting blocks).
5. Use the mouse to select a single block which the program uses to calculate a flat baseline. Alternately, you can click on a single point (instead of selecting a block).
6. Keyboard entry of beginning and ending baseline values.
7. Keyboard entry of a constant baseline value.
For all methods except #2 (automatic periodic references) or #3 (multiple blocks), the baseline is drawn on-screen (if it fits within the current Y-axis scale) and you can cancel, reselect, or select it. The final baseline is drawn across the entire file.
See the Reference removal section for an easy method of eliminating baseline sections for analysis.
The main goal of lag correction is usually to correctly synchronize two channels. To determine how much lag time to use, follow these steps:
NOTE: Because numbers are stored in floating point format with 8-10 digit precision, integration may result in loss of least-significant digits, particularly if the numbers are large (i.e., if their cumulative sum contains more than 8-10 decimal digits). Thus integration followed by derivation of the same data will not necessarily produce a result identical to the original data. Normally this poses no problems to users. |
The program makes a memory copy of the channel
and uses it as its data source (this insures that within a given smoothing
iteration, previously smoothed data are not part of the calculations).
When smoothing is complete, the channel is redrawn. If a block has
been selected, you can smooth either all the data or just within the block.
You can also smooth conditionally, changing only those data within a certain
range (i.e., greater than or less than user- specified values). If
the file contains less than 24 channels, you can copy the active channel
to a new channel before smoothing.
The example at right shows 19-sample smoothing to be repeated over six cycles. Notice that the indicated smoothing interval (calculated from the sample interval and the averaging interval of 19 samples) is 27 seconds. Conditional smoothing is selected, and smoothing will be applied only to data points will values less than 0.15.
Smoothing can be repeated as often as necessary, but be very careful that you do not obliterate useful information from the data. For example, smoothing a high-frequency waveform may result in a substantial decrease in the peak amplitude, although this should not affect frequency calculations. Remember that you can undo only the last smoothing operation performed.
Two options are available for sweeping: 'Single scan' makes a single pass through the data; 'Spike sweep' repeats scanning until all appropriate spikes are removed.
The trim data option searches for data points that are greater
or less than preset limits; all data outside these limits are set to user-defined
values. Both sweep and trim can be performed within
blocks only (if a block has been selected).
The Absolute Value option converts all data to its absolute value (positive or zero); this is also available in the transforms window.
In this example, the user has performed a spike sweep with a threshold value of .02, and the program found and removed 78 spikes. Also, the trim data options have been set such than any value exceeding 5 will be set to equal 5, and any value less than -.15 will be set to zero.
Interactive: In Spike mode, spikes are selected by clicking on them with the crosshair cursor. The program replaces the selected point with the average of the two values on either side. This is most effective for spikes of short duration (optimally, a single sample).
An alternate noise-removal method is 'Data
Linearizing'. This mode linearizes all data points between two
selected points (as shown schematically in the example plot). Points can be selected by single clicks, or by the standard click-hold-drag method. Optionally
(as shown here), you can select the 'Interpolation markers' option
to automatically insert the standard double-arrowhead interpolation markers
at the beginning ("»") and end ("«") of
linearized intervals (this lets you avoid using interpolated data in most
analyses).
If you want to conserve the short-term variation in the data within the 'linearized' area, select the 'Conserve variation' option. This uses a 3-point nearest-neighbor smoothing operation to reduce the amount of deflection within the defined area, but it preserves the point-to-point variation. You can select how much linearizing will occur with the pop-up menu of smoothing repetitions ('smooth once' = 1 smoothing pass, i.e., the most variance preseved, and 'smooth X5' is 5 smoothing passes, resulting in the greatest degree of linearizing). As more smoothing passes are added, the corrected data increasingly resemble a straight line between the beginning and ending points (this is best visualized by experimenting with the three options, using the 'delete' key after each one to eliminate the correction).
You can alternate between spike removal mode and linearizing mode by first clicking in the lower left window, then clicking the desired mode button, and finally returning to the plot area. Accessing the 'Quit' button is done similarly.
If you are in multi-channel display mode, the program will shift to single-channel display (of the currently active channel) for spike removal operations. If the file contains more than screenwidth cases and you are in Entire File display mode, the program will shift to Active Screen mode when interactive spike removal is selected. Use the screen slider to shift between screenfulls of data (note that the data linearizing option can work across screen boundaries).
The Remove references window is shown above. To use it you need to specify:
Reference removal is performed and the new values are displayed in the plot area when you click the 'Fix references' button. You can then undo the results, or exit the function.
This sample window shows the addition of two channels, with the result stored in a new channel. The multiply/divide option is set to 1 (meaning that no action is taken on the result of the addition operation). If you set this number to something other than 1.0 and click either of the multiply or divide buttons, the result of the 'main' operation will be adjusted accordingly.
Two buttons at left toggle between transformations and unit conversions. The operations for transformations are selected from the smaller (right-most) of the two pop-up menus shown at near right. Note that the last five transformations (addition, subtraction, multiplication, and division with different channels, and Q10 correction from a temperature channel) are only available if the file contains more than one channel.
Alternately, you can use a range of unit conversions from the pop-up menu shown at near right. These pertain to commonly-used units of energy, flow rate, pressure, mass, speed, temperature, water vapor pressure, gas volume, and so forth. Conversion into mass-specific units is also available. The 'STP converter' option does not affect the channel data, but allows you to calculate an appropriate correction for temperature and pressure effects on gas volume (useful in respirometry).
For most of these operations, the conversion factors (i.e., the numerical coefficients in the conversion equations) can be edited as desired prior to calculations. For example, you may wish to adjust the conversion factor for the heat of vaporization of water (which is set for a default evaporative surface temperature of 37 °C) to a value accurate at some other temperature.
Many of the conversions will attempt to adjust the channel label as appropriate. For example, if you have chosen to convert a channel to mass-specific units, the program will append a "/g" or "/Kg" to the end of the current label. Similarly, if you are converting a channel with units of watts to units of KJ/day, the program will search the label for the word "watts" and replace it with "KJ/day". Keep in mind that automatic label modification is not always successful (depending on the format of your channel labels), so be sure to check.
Note that the transformations routines do not make any attempt to adjust labels to reflect changes. You should be diligent in correcting the file labels appropriately in order to avoid confusion later on.
During mathematically complex operations that are time-consuming (such as calculating logs or computing vapor pressure from temperature), a progress bar appears. This isn't shown for faster operations, since drawing the progress bar often is slower than the computations themselves!
Some special considerations for transformations and conversions:
The response correction option is similar to the 'effective volume' computation used for gas exchange. It compensates for capacitance-like characteristics that slow response times and obscure rapid changes in measured variables, using the "Z-transformation". The algorithm compares successive values and corrects them according to the following equation:
corrected value = [(value-last value)/Z factor] + last value
Typical Z factors are between 0.1 and 1.0 (other values are accepted). It is reasonable to determine the correct value by trial-and-error, if you have a recording that contains a known step change (or near-instantaneous change). Apply different Z factors until the results approximate a step change.
The breakpoint transforms option lets you use different transformation equations depending on whether data are less than or greater than a user-specified 'breakpoint' value. You enter the equations and the breakpoint in the following window (which appears whenever you select the breakpoint option):
After entering the breakpoint value, pick the type of equation for data below the breakpoint value, using the radio buttons at left. Then click the 'use this equation for data < breakpoint' button, and enter the a, b, and c values in the first set of edit fields at the bottom of the window. Then select the equation type for data equal to or greater than breakpoint, click the 'use this equation for data > breakpoint' button, and enter the a, b, and c values in the second set of edit fields. In this example, a 2-way polynomial will be applied to data < breakpoint, and an exponential equation will be applied to data > breakpoint. When all is correct, click the 'OK to transform' button to return to the main Transformations and Conversions window.
NOTE: LabAnalyst X does not allow invalid or meaningless mathematical operations, such as division by zero, taking the log of a negative number, or using a non-integer exponent on a negative number. If LabAnalyst X encounters these situations during a transformation, the offending cases are set to zero, the computer beeps, and a warning message appears. At this point you can click the 'Cancel' button to discard the transformed data, or accept the (partially invalid) transformation.
Custom transformations can be accessed, edited, and deleted from the EDIT menu, but can only be created from within the standard transformations window. To do this, set up the transformation as you would normally (i.e., select the mathematical operation and supply the correct coefficients). Then select ‘Add new custom transform’ from the ‘Custom transformations’ popup. A window will open asking for a name for the new transformation; you can enter anything you like, up to a maximum of 40 characters.After the transform name has been entered, you should select ‘Manage custom transforms’ (again, from the ‘Custom transformations’ popup) and see if the new item appears in the list. Each new transform is appended to the end of the existing list of custom transformations. From this window you can click buttons to delete one or more transforms (caution: this cannot be ‘undone’) or to edit the parameters of a transform. In the latter option, you can change numeric coefficients but not the basic mathematical operation. Make sure you don’t enter coefficients that will not ‘work’ with the type of transform (e.g., don’t try to take the square root of a negative number).
• Using custom transforms: To use a custom transformation, go to the standard ‘Transformations’ menu option, select ‘Manage custom transforms’, select the desired transform from the list in the new window, and click ‘Selection OK’. The ‘Transformations’ window will show the name of your selection. Pick the appropriate channel(s) and then click the ‘Transform all’ button, as usual.
• NOTE: At present, the custom transformations options do not work for ‘breakpoint’ conversions.
For related operations, see also the SCALE RESULTS option (ANALYZE menu).
Results are stored in the first of the source channels, or optionally (if the number of channels is less than 24) in a new channel (edit the file label as appropriate). If a data block has been selected, you can transform all the data or just within the block. A typical expression transforms window looks like this:
Some general considerations:
NOTE: This will only 'catch' errors in the basic numeric expression. It may not detect invalid or meaningless math operations that may be attempted when channel data are processed, such as division by zero, or taking the log or a non-integer exponent of a negative number. If such situations occur, results may be unpredicatable. The algorithm does find most such errors during processing, however.
The underlying code for the expression evaluator was largely developed by Robert Purves (recently deceased and greatly missed). I borrowed it -- with his permission -- and made some modifications for LabAnalyst. But Robert P. deserves all the credit.
Resets the scale and range of the data to values specified by the user, based on a block of data. For example, you can rescale a block of data with an initial range of -.134 to .335 to a new range of zero to 100% (or any other upper and lower limits).
Note: although calculations are based on data within the selected block, the scale and range adjustment is applied to the entire file.
If you use the 'Block only' option, the rescaling is applied only to the selected block -- but keep in mind that it will now be out of calibration with the rest of the data in the channel.
BLOCK SCALE, 0-100% This option automatically scales the selected block to a range of zero to 100.
NOTE: although the plot area on the screen shows the compacted version, any analyses are performed on the original data, which is shown as normal in the block window.
This example shows a 20-fold compaction, which equates to a 'block size' of 30 seconds for each mean and S.D. computation. The 'highlight data means' and 'show S.D.' options are selected; this will show the mean values as a dot, and the SDs as vertical lines in the plot area (in active screen mode only if there are more than screenwidth cases).
Some operations, like Baseline with user-selected points, are difficult to use if the data are shown in compacted form (although you can still read the uncompacted values for the cursor position in the data bar). It's probably best to shift to the regular display ('normal plot') for these operations.
CHANGE BLOCK This option allows you to replace the values within a selected block with a user-defined constant. Optionally, you can do this to multiple blocks simultaneously. This is useful in many situations. One example is if you wish to apply the 'minimum value' analysis operation, but your data contains some reference readings. By marking the references and then changing them to some arbitrary large value, you can scan the complete data set and the analysis will be restricted to 'real' data instead of being confused by reference points.
![]()
REMOVE BLOCK This option removes the data within a marked block, either within a single channel or for all the channels (the 'entire file' option). In either case, the elapsed time and time-of-day readouts in the data bar are no longer accurate for data points later in the file than the removed segment. The one exception to this rule is if you deleted a block at the beginning of the file (i.e., starting at the first case).
REMOVE CHANNEL K This option removes one or more entire channels from the file. At least one channel must remain. The 'undo' option cannot restore a channel or channels after they are removed from the file.
A check-button lets you opt to have a suffix indicating re-arrangement appended to the file name.
NOTE: This operation is not undo-able.
Other links: |