Scanning Transmission Electron Microscopy Facility
PCMass Manual: Tutorial 5
|
Global Statistics
|
- Box up the magnification bars for images from one specimen as in Tutorial 4. Summaries will be saved in the SpecAves folder
with names U*****_aves.txt and U*****_parts.txt, where U*****is the specimen number. Therefore it is important in the summaries
to include images from only a single specimen.
- Open any image in the sequence, strike <o> and select the model and type of measurement
(usually “align pass”) for which you want a global histogram. Hold down the <Shift> key and strike
<o>. If you want a global histogram for additional models, repeat the selection step and
<Shift><o>. The summaries will be appended to the same file (with duplicated global statistics).
If you want additional histograms, select a different model and repeat <SHIFT><o>.
The additional information will be appended to the U*****_aves.txt and U*****_parts.txt files.
- View the U*****_aves.txt and U*****_parts.txt files with a text editor such as Notepad. Use a non-proportional font to keep
the columns straight. Delete any redundant information. The screen display is useful for up to 16 files. Note the numbers of
the “best” (at least 3 good TMV segments with SD<3%, some particles/filament segments of interest) and “worst” few images.
- If you want any of the screen information saved, hold down the <Alt> key and hit
<PrintScrn>. This copies the entire screen to the clipboard. Open PowerPoint, PhotoShop or Paint,
select new and paste. This should give a copy of the screen which you can carve up and save as a *.jpg. You can do the same
with SnagIt or other screen copy programs.
- Check the "best" image files by opening them and judging the quality of the selection with the <i>
and <o> modes. Pay particular attention to "outliers" and objects picked up by more than one model.
Only delete "outliers" (<Delete> key in <i> mode) if they have an obvious problem
(end of a filament, obvious salt, etc.). It is better to use an objective criterion based on selection parameters. Missed objects
can be added using the manual mode, followed by <F12>. That will only realign particles not previously done.
- Check the "worst" images in order to understand what went wrong (overcrowded, too sparse, salty, etc.). If they cannot be
salvaged by tuning up models or selection parameters or by deleting defectives, it is best to delete the entire *.smm file for
that image.
- Repeat steps 2, 3, 5 & 6 until you are satisfied that the classifications are correct and defective objects are being excluded
by choice of selection parameters or deletion. Erase the U*****_aves.txt and U*****_parts.txt files and re-run <SHIFT><o>.
- At the end of the aves.text file you should find a summary of those files with good TMV. For each such file the particle mass is
scaled to the known TMV M/L of 13.1 kDa/A. and a new histogram is formed using scaled mass values.
- The overall accuracy is difficult to estimate due to the possibility of systematic errors. Some people suggest using the standard
error of the mean (the SD/sqrt(number of measurements)), but this is overly optimistic. We recommend instead using the SD of the means
of the individual images. However, careful attention should be paid to the possibility of systematic errors as indicated by significant
deviation from 13.1 in TMV M/L or significant variation in background from image to image.
|
Last Modified: June 12, 2009
Please forward all questions about this site to:
Denise Monteleone
|
