Boulder Laboratory for 3-Dimensional Electron Microscopy of Cells MTPAIRING(1) MTPAIRING(1) NAME mtpairing - to analyze pairing between MTs SYNOPSIS mtpairing DESCRIPTION MTPAIRING calculates the length over which MTs are paired with each other in 3-D, and allows one to recolor MTs absed on these pairing lengths. It can also assign polarities to MTs based on the positions of their endpoints in Z, and allows one to recolor MTs based on these polarities. The program combines features of MTOVERLAP and GENHSTPLT. It also has several exploratory features that are not documented because they were not particularly useful. MTs are considered "paired" when they are within a certain distance of each other in the X/Y plane. Typically, you would set this distance to be the upper limit of the peak from a neighbor density analysis. The absolute pairing length for a pair of MTs is the total length in Z over which they are within that distance of each other. A fractional pairing length is also computed; this is the pairing length divided by the length over which the two MTs appear in the same sections. Before running the program, you must figure out how to specify which MT's are in a bundle. If all of the MT's in a model belong to one bundle, then this task is easy. If you have several bundles in one model, then you have several alternatives. One is to determine the lower and upper X, Y and Z coordinates of a box, such that the bundle consists of all MT's that contain at least one point within the box. Another way is to make a model object within the plane of one section to serve as a boundary contour. This contour, together with a lower and upper Z coordinate, specifies a "cylinder", and this program will include in the bundle any MT with at least one point inside this cylinder. The most elaborate way is to make a series of model objects for boundary contours in different sections. The program will then include in the bundle any MT that is included within any one of the contours. When you enter X, Y or Z coordinates for this purpose, they must be index coordinates of the image file. That is, X and Y values must be in terms of pixel coordinates, and Z values must be in units of the original section numbers, before adjustment for tilt or scaling by section thickness. If the sections were significantly tilted during microscopy, the program can adjust for these tilts given the proper information. Prepare a file in which the first line shows the Z value and the tilt of the first tilted section (or of the first section, if that one was tilted), and each successive line shows the Z value and tilt for each section on which tilt was changed. Z values should occur in ascending order. When you start the program, you will have to make a standard series of entries until you get the first display. From there, you can select a number of options to loop back and change those entries. Initial entries in order are: 0 for plots in the graphics window, or 1 for plots only on the terminal. Note that if you need to use terminal plots, you will need to specify that option each time that you do a plot. Name of command file to take entries from, or Return to continue making entries from the keyboard. The program can read entries from a file instead of from the keyboard, then switch back to keyboard input if the file ends with the appropriate entry. Number of bundles to read from model files, or 0 if the entries specifying all of the bundles are in yet another file. IF you enter a positive number, then enter for each bundle: Name of model file with bundle in it, or Return to use same file as previous bundle IF you enter the name of file, then make the following 3 entries: Name of file with information on tilt angles, or Return if there is no such file (pictures taken at 0 tilt) Section thickness in nm, to scale Z coordinates to microns; or / to leave Z values unscaled Magnification of negatives, and scale of digitization (the value of microns/pixel from VIDS), to scale the X/Y coordinates correctly; or / to leave X/Y coordinates unscaled. This entry makes no difference unless you choose to calculate one of the special three-dimensional overlap factors. Number of limiting regions (boundary contours or rectangles defined by X/Y coordinates) needed to specify the bundle, or 0 to take all of the objects in the model. For each limiting region, then enter: The number of an object specifying a boundary contour, or 0 to enter limiting X and Y coordinates of a box. IF you entered 0, next enter the lower and upper X index coordinates and the lower and upper Y coordinates of the box, or enter / to have no limit on the X and Y coordinates THEN enter the lower and upper Z coordinates of the box (in units of sections), or / to have no limits on Z coordinates IF you entered the number of a boundary object, next enter lower and upper Z coordinates of the "cylinder", or / to set those limiting coordinates to the Z coordinate of the boundary contour. The latter is typical if one uses several contours in different sections to specify the bundle. IF you entered 0 for the number of bundles, next enter instead the name of a file. The first line of this file should have the number of bundles specified there. The rest of the file should be all of the entries just described for each bundle. Enter a list of numbers of the bundles to work with. Ranges may be entered, e.g. 1-3,7-9. The lower and upper limits of Z within which to compute pairing. A minimum number of sections to assume as shared sections when the fractional pairing is computed. This entry was intended to avoid unreasonably large fractional pairing lengths when two MTs only appear together in a few sections. A value of 4 may be useful. Enter a list of the types (colors) of MTs for which to compute pairing. These will be the "reference MTs" in the pairing calculations. Type Return to include all MTs. Enter a list of the types (colors) of MTs to consider as neighbors to those reference MTs. Type Return to include all MTs. Enter 1 for a simple pairing factor which is 1 for MTs within a certain distance of each other and 0 beyond, or 2 or 3 for a pairing factor that decays with distance in the X/Y plane, either as an inverse power or exponentially. Enter the distance in the X/Y plane at and below which overlap will equal 1. The distance should be in microns if you have scaled X/Y values, or in pixels if you have not. IF you entered 2, next enter the power for the decay (e.g., with a power of 2, overlap will decay as the inverse square of distance) IF you entered 3, enter instead the space constant for exponential decay. Overlap will be 1/e less for MT's separated by 2 space constants than for MT's separated by 1 space constant. Distance should be in microns if you have scaled X/Y values, or in pixels if you have not. Minimum pairing length that a pair of MTs should have before its data will be stored for examination. If there are not hundreds of MTs, a minimum of 0 will retain all data about MTs with any pairing. At this point, the program computes the pairings and gives information about the lengths of the many pairs of MTs with no pairing. You are then at the option point. Options are: 1: to plot the pairing data about each MT. Columns available are: 1 = MT length 2 = absolute pairing length summed over all neighbors to the MT 3 = maximum pairing length achieved with one other MT 4 = fractional pairing length summed over all neighbors to the MT 5 = maximum fractional pairing length achieved with one other MT 6 = Z value of midpoint of MT. 2: to plot the data about paired MTs. Columns available are: 1 = arithmetic mean of the lengths of the two MTs 2 = geometric mean of the lengths of the two MTs 3 = absolute pairing length of that pair 4 = fractional pairing length 9 = mean separation between MTs while they were paired 10 = SD of separation 11 = coefficient of variation = SD/mean of separation With either option 1 or 2, you must enter the numbers of the columns to be plotted on the X or Y axes. Next, enter a number for the symbol type as commonly referred to in Genhstplt and other places. After this, you will enter the BSPLT subroutine, whose entries are described in the manpage for Bsplt. 3: to loop back to the point where you specify which bundles to work with and then enter other parameters of the pairing calculations. 4: to loop back and read in new bundles, replacing existing ones. 5: to loop back and read in new bundles, retaining existing ones. 6/7: to plot the current Postscript file on the screen/printer 8: to exit the program 9: to recolor the model. After selecting this option, you make an indefinite series of entries of the following form. In one line, you enter the following information to select a set of MTs: New color of MTs. Enter -1 here to terminate the series of recolorings. Column to use to select MTs. The data about each MT are referred to by positive column numbers (1 to 6 as described above); the data about pairs are referred to by negative column numbers (-1 to -4 as described above). An entry of 0 will use the "polarity" values determined after using options 11 and 12. The lower and upper criterion limits to apply to values in that column. 0 to select MTs that are within the limits, or 1 to select ones outside the limits. On the next line, enter a list of the original colors that MTs should have in order for them to be recolored according to these criteria, or Return to apply the criteria to MTs of all colors. In this way, you can recolor each color of MT that meets a particular criterion to a particular new color. After you have entered all of the selections, enter the name of the output file in which to place the recolored model. 10: to take commands from a file (next enter filename, or Return to take input from the keyboard) 11 will find clusters of mutually paired MTs (which can be all of the MTs in a bundle), 12 will find polarities based on positions of the MTs in the bundle or cluster, and 13 will graph the clusters. These features are not documented here - consult command files for examples (these files may call this program "COMSYMP"). HISTORY Written by David Mastronarde, 1993