User vs. software-dependent variability of ELISPOT counts obtained from ten different laboratories

Abstract

Application of microscopy to evaluate the morphology and size of filamentous proteins and amyloids requires new and creative approaches to simplify and automate the image processing. The estimation of mean values of fibrils diameter, length and bending stiffness on micrographs is a major challenge. For this purpose we developed an open-source FibrilJ plugin for the ImageJ/FiJi program. It automatically recognizes the fibrils on the surface of a mica, silicon, gold or formvar film and further analyzes them to calculate the distribution of fibrils by diameters, lengths and persistence lengths. The plugin has been validated by the processing of TEM images of fibrils formed by Sup35NM yeast protein and artificially created images of rod-shape objects with predefined parameters. Novel data obtained by SEM for Sup35NM protein fibrils immobilized on silicon and gold substrates are also presented and analyzed. Program summary Program title: FibrilJ Program Files doi: http://dx.doi.org/10.17632/ndxb93h4vc.1 Licensing provisions: Apache-2.0 Programming language: ImageJ Macro Language Nature of problem: Amyloids are large protein aggregates that form unbranched fibrils. Formation of amyloids by different proteins leads to the emergence of a number of serious human diseases, including Alzheimer's, Parkinson's and Huntington's diseases. Therefore, various amyloids, in particular their topology, are widely studied in an enormous number of papers. The structural organization of the amyloid aggregates at the molecular level is a key problem in this area. At the moment new methods of amyloids analysis based on electronic and probe microscopy are particularly popular. Usually it is required to process images containing hundreds of amyloids to obtain statistically reliable data on the distribution of the thickness, length and persistence length (a basic mechanical property quantifying the stiffness of a polymer). The lattermost parameter was introduced for the characteristic of bending stiffness of amyloids similarly to that in a model of wormlike chain for polymers. The manual measurements of these parameters take a lot of time and, due to the human factor, are insufficiently representative. ImageJ/FiJi is open-source software and one of the most commonly used program for image processing. The ImageJ/FiJi community provides BoneJ plugin for bones analysis and DiameterJ plugin for nanofibers analysis, which have the function of determining the diameter of the respective objects. Also, approaches that involve calculations of the end-to-end distances and contour lengths of biopolymers were used to determine their persistence lengths using ImageJ/FiJi. These above mentioned plugins are not designed for the study of amyloids and have never been used or validated for them. They do not provide the necessary functionality for the automatic recognition of amyloids on the substrate and filtering objects by a number of parameters. They have unspecified accuracy of these objects measurements and do not calculate the objects distribution on average diameters and persistence lengths. Solution method: For this purpose we developed a FibrilJ plugin for ImageJ/FiJi program with an open source code. It automatically recognizes the fibrils on the surface of a mica, silicon, gold or formvar film and further analyzes them to calculate the distribution of fibrils on average diameters, lengths and persistence lengths. FibrilJ usage may unify the process of diameter and persistence length calculation and it may also contribute to comparison of results obtained by different research groups. Plugin FibrilJ which is presentedin this paper is free of DiameterJ or BoneJ weaknesses. It has been validated by artificially created images of rod-shape objects with predefined parameters and micrographs of Sup35NM protein fibrils.