Boulder Laboratory for 3-Dimensional Electron Microscopy of Cells
HOWFLARED(1) HOWFLARED(1)
NAME
howflared - measure microtubule end flaring and curvature
SYNOPSIS
howflared [options] output_file
DESCRIPTION
Howflared analyzes tracings of the walls of microtubules (MTs) to compute
various measures of the amount of curvature or "flaring" at the MT end. The
IMOD model with MT tracings must meet a number of requirements. The first
step is to extract the MTs into subvolumes so that they are oriented
vertically, with the end facing downward. This can be done with
Mtrotlong. The orientations do not need to be perfect. Once subvolumes
have been extracted as a set, they can be loaded into 3dmod together as a
4-dimensional data set. The 4th dimension is referred to as "time".
Alternatively, a central slice can be extracted from each and a stack made
with these central slices.
The tracings must all be drawn starting near the top of the image in Y and
ending at the MT end. Each tracing defines two regions of an MT wall: a
linear segment before the end, and a potentially curved segment. A line is
fit to a defined part of the linear segment and extrapolated into the region
of the end. The quantitative measures are based on the deviation of the
curved segment from this extrapolated line. Thus, a tracing must include a
linear segment, and there must be some way to indicate where the linear
segment ends and the measurement of the deviation should begin. There are
essentially three different ways to indicate this boundary.
One way is to specify how far down from the top of the tracing the
beginning and the end of the line fit should be. This method is suitable if
all the tracings have similar starting points near the top of the subvolumes
and their curvature starts at similar locations as well. The starting and
ending points can be specified as either absolute units (in pixels) or as a
fraction of the distance from the maximum to the minimum Y value in the
tracing. The second method is to use a horizontal line, consisting of only
two points, entered as a separate contour, to specify the boundary point.
With this method, one might still enter an ending limit, in which case the
maximum of the ending limit and the Y position of the marker line are used
for the end of the linear fit, but the measurement of deviation starts at
the marker position. The third method is to model the linear segment with
just two points. In this case, one would enter 0,0 for starting and ending
limits. The second and third methods can both used in the same model; the
second point will be used as the boundary only if there is no horizontal
marker line.
The program can analyze either two paired walls of a MT, or a single wall.
When two walls are paired, the linear fit finds a single slope for the two
walls, and a width between them. This width is used for normalizing some of
the flaring measurements. If there are also single, unpaired walls, a width
is assumed when normalizing these measurements (see -width option). When a
horizontal marker is used for a pair of walls, it should be drawn from near
the boundary point on the left wall to near the boundary point on the right
wall. A single wall is treated like the wall on the left of a pair; so in
this case a marker line should have its left end near the boundary point on
the wall.
The program needs to be able to tell when two MTs are paired, and when a
marker line belongs with one or two MTs. Nothing special needs to be done
when modeling in only a single plane per MT. If you are modeling in
multiple sub-volumes, the program will segregate the tracing based on their
times; if you are modeling in a stack of single slices from different MTs,
the segregation will be based upon Z value. However, if you want to trace
multiple, unpaired walls in the Slicer, either you need to segregate them by
placing them in either separate objects or separate surfaces, or you need to
use the -nopairs option. Thus, you could have one object for all of the
first tracings in each MT, a second object for the second tracing, etc.
Similarly, you could use the surface number 0 for the first tracing in each
MT, 1 for the second, etc.; or you could simply put each tracing (or tracing
pair) with associated marker into a separate surface. If you are willing to
have no marker lines, then you can use the -nopairs option to indicate that
there are neither pairs nor markers.
The program computes various quantities and stores them in columns of a
data matrix; you specify which columns you want output with -columns. The
items and their units are:
The area between the tracing and the extrapolation of the wall, down to
the minimum Y value of the tracing (square nm).
The square root of this area (nm)
The sum, over all line segments, of the product of the line segment length
and the angle between the segment and the extrapolation of the
wall(nm*radians)
The area divided by the square of the average MT width (square MT widths)
The square of the area divided by the average MT width (MT widths)
The angular sum divided by the average MT width (MT widths * radians)
The total length of the tracing after the boundary (nm)
The final angular deviation of the tracing from the extrapolation of the
wall (degrees)
The angular change in degrees per nm of length
The average radius of curvature in nm.
The first 6 items are placed into columns 1-6 for the left wall of a pair
or for an unpaired wall, and into columns 7-12 for the right wall of a pair.
The sum of the two values for left and right in placed in columns 13-18.
The last 4 items are placed into columns 19-22 for the left wall of a pair
or for an unpaired wall, and into columns 23-26 for the right wall of a
pair.
In the output file, these columns of output are preceded by 3 or 4
numbers. The first number is a model identifier; the second is the object
number. If you used surface numbers to segregate tracings, the third of 4
numbers is the surface number; if you used the -nopairs option, the third of
4 numbers is the contour number. The last number is the time value, or the
Z value.
Howflared uses the PIP package for input exclusively (see the manual page
for pip). The following options can be specified either as command line
arguments (with the -) or one per line in a command file or parameter file
(without the -):
-output OR -OutputFile File name
Output file for results. If this option is not entered, the first
non-option argument will be used for the filename.
-columns OR -ColumnsToOutput List of integer ranges
List of columns to output (values ranging from 1 to 26). For paired MT
walls, columns 1-6 and 19-22 are for the left walls, 7-12 and 23-26 are
for the right walls, and 13-18 are the sum of values from 1-6 and 7-12.
Unpaired walls are treated as left walls, and their values will appear in
both columns 1-6 and 13-18. See description of computed values above.
-point OR -PointOutputFile File name
Output file for points in each wall, starting at the point where the flare
measurement begins. The points are rotated so that the wall above that
point would be vertical, and they are expressed in nanometers relative to
the starting point. For each MT wall, there is a line starting with the
number of points, then containing the 3 or 4 numbers described above as
preceding the columns of output in the standard output file, and ending
with 1 for a left or unpaired wall, or 2 for a right wall.
-pixel OR -PixelSizeDefault Floating point
Default pixel size in nm, which will be used for any models that do not
have a pixel size defined. The default is 1, namely results will be in
pixels.
-width OR -WidthDefault Floating point
Default width between walls in nm. This value is used to compute
normalized areas for unpaired walls. The default is 20.
-surface OR -UseSurfaceNumbers
Use surface numbers to determine whether two wall tracings are paired and
whether a horizontal marker contour matches a wall tracing. This is not
the default because surface numbers might have been introduced by
accident.
-nopairs OR -NoPairsOrMarkers
There are no paired tracings and no horizontal marker lines. With this
option you do not need to segregate tracings in separate objects or
surfaces.
-model OR -ModelFile File name
Input model file with tracings of microtubule walls
(Successive entries accumulate)
-fit OR -FitTopAndBottom Two floats
Upper and lower limits for the line fit. The limits are entered as either
an absolute distance from the maximum Y for a MT or a fraction of the
distance from the maximum to the minimum Y. If the upper limit is 0, the
maximum Y is used. If the lower limit is 0, the limit used depends on
whether there is a horizontal marker line or not. If there is a line, it
is used as the limit; if not, the second point of the contour is used as
the lower limit. Enter this option once to use the same limit for all
models, or once per model. If it is not entered, the default limits are
0,0.
(Successive entries accumulate)
-id OR -Identifier Integer
ID number for output. Enter this option either once to specify a starting
ID number, or once per model. If it is not entered, the ID numbers start
at 1.
(Successive entries accumulate)
-help OR -usage
Print help output
-StandardInput
Read parameter entries from standard input.
HISTORY
Written by David Mastronarde in 1996
Modified for PIP input and incorporation into IMOD, 10/14/06