AFM provides a fast and inexpensive method to characterize amyloid morphology down to single monomers (a single protein molecule). The robustness of the aggregates also makes it an ideal sample since the drying process does not cause structural changes. This allows us to rapidly prepare and screen multiple samples and make quantitative nanoscale measurements. The automated segmentation and extraction of statistical data is possible, though more demanding due to the polymorphic structure of amyloids.

The major prion protein (PrP) is a prominent example, causing neurogenerative diseases in many species such as bovine spongiform encephalopathy (BSE, mad cow disease),  scrapie, and  Creutzfeldt–Jakob disease in humans (CJD). Individual PrP fibrils below are approximately 10 nm in diameter and aggregate into micron-size structures.

PrP deposited on mica and dried, showing a wide range of aggregate sizes.

Visualization of PrP fibril structures.

Automated segmentation of PrP amyloids in a 2×2 um scan to extract statistical quantities. Segmentation of such polymorphic and polymeric structures is more complex than single particles and the resulting statistical data is more difficult to interpret.

Automated segmentation result.

Extracted PrP amyloid height distribution.

Extracted PrP amyloid volume distribution.

Extracted PrP amyloid surface area distribution.

Extracted PrP amyloid major axis length distribution (amyloid shape).

Extracted PrP amyloid minor axis length distribution (amyloid shape).

The software also outputs a segmented image with labeled segments and a corresponding table with the extracted properties of each segment. 

Example of segment-level data comparing segments 37, 46, 51, 53 and 57 in the center of the image.

Note that the lateral dimensions are influenced by AFM tip convolution.