Several automated crystallization systems have recently been developed for high-throughput X-ray structure analysis. However, the evaluation process for the growth state of crystallizing protein droplets has not yet been completely automated. This paper proposes a new evaluation method for crystalline objects in automated crystallization experiments. The main objective is to determine whether a droplet contains crystals suitable for diffraction experiments and analysis. The evaluation method developed here involves extracting line-segment features from an image of the droplet and discriminating the state of crystallization using classifiers based on line features. In order to verify the efficacy of the proposed method, it was used to classify images obtained by an automated crystallization system.

In a usual crystallization process, the researchers evaluate the protein crystallization growth states based on visual impressions and repeatedly assign scores throughout the growth process. Although the development of crystallization robotic systems has generally realised the automation of the setup and storage of crystallization samples, evaluation of crystallization states has not yet been completely automated. The method presented here attempts to categorize individual crystallization droplet images into five classes using multiple classifiers. In particular, linear and nonlinear classifiers are utilized. The algorithm is comprised of pre-processing, feature extraction from images using texture analysis and a categorization process using linear discriminant analysis (LDA) and support vector machine (SVM). The performance of this method has been evaluated by comparing the results obtained using the method with the results obtained by a human expert and the concordance rate was 84.4%.

PurposeThe purpose of this paper is to present classification schemes for the crystallization state of proteins utilizing image processing.Design/methodology/approachTwo classification schemes shown here are combined sequentially.FindingsThe correct ratio of experimental result using the method presented here is approximately 70 per cent.Originality/valueThe paper is a contribution to automated evaluation crystal growth, combining two classifiers based on specific visual feature, sequentially.

No abstract

PurposeThe purpose of this paper is to propose a state discrimination for crystallization samples (droplets), the purpose of which is to discriminate between diffractable extracts (crystal) and other objects.Design/methodology/approachThe line feature from the image of the protein droplet was extracted and the state discriminated using a classifier based on line features. A support vector machine is used as the classifier.FindingsIn order to verify the performance of the proposed method, the growth state was discriminated experimentally using the images taken by TERA, an automated crystallization system. The correction ratio was determined to exceed 80 percent.Originality/valueContribution to automated evaluation process of the growth state of protein crystallization samples.