• Use a cell line that adheres well to the rubber substrate, but is known to have a diffuse IF network. Keratin IFs tend to bundle, so try to avoid epithelial cells.
• To avoid stress-relaxation, strain cells as fast as possible, and fix as fast as possible.
• Use modified dehydration and embedding protocol and make blocks.
• Should also do some cells that are extracted (see protocol), so that only IF network remains. This could make clearer pictures and will also increase the number of filament images that are usable for analysis.
• When cutting sections, it is vital that the direction of stretch be taken into account so that IFs can be photographed in the proper orientation. We will want to note the direction of stretch on the published images, and we may also want to analyze filament diameters as a function of orientation to the stretch.

Analysis of Filament Images

• Open images in ImageJ
• Measure the average background pixel intensity for each filament, both sides
• Straighten filament using Straighten plugin, 250 pixels wide
• Flip image, save straightened image
• Measure plot profile, paste into Excel spreadsheet
• Normalize data (0-100 range) using background as minimum, max value as max
• Calculate symmetry of plot profile using percent comparisons for three pixel ranges
• Exclude data that do not fit symmetry criteria
• Calculate diameter for each filament using a pixel intensity threshold (85%, 90%?)
• Flip remaining curve over (subtract threshold value, flip sign)
• Integrate over curve by adding up total pixel intensity
• This will be a better measure of the relative distribution of mass in a given cross-section
• than just a single diameter taken by adding up the number of pixels above the cutoff.