Computational models in the service of X-ray and cryo-EM structure determination

CASP (Critical Assessment of Structure prediction) conducts community experiments to determine the state of the art in computing protein structure from amino acid sequence. The process relies on the experimental community providing information about not yet public or about to be solved structures, for use as targets. For some targets, the experimental structure is not solved in time for use in CASP. Calculated structure accuracy improved dramatically in this round, implying that models should now be much more useful for resolving many sorts of experimental difficulty. To test this, selected models for seven unsolved targets were provided to the experimental groups. These models were from the AlphaFold2 group, who overall submitted the most accurate predictions in CASP14. Four targets were solved with the aid of the models, and, additionally, the structure of an already solved target was improved. An a-posteriori analysis showed that in some cases models from other groups would also be effective. This paper provides accounts of the successful application of models to structure determination, including molecular replacement for X-ray crystallography, backbone tracing and sequence positioning in a Cryo-EM structure, and correction of local features. The results suggest that in future there will be greatly increased synergy between computational and experimental approaches to structure determination.

Multiple mental representations in picture processing

This article investigates whether goal-directed learning of pictures leads to multiple mental representations which are differently useful for different purposes. The paper further investigates the effects of prompts on picture processing. 136 undergraduate students were presented maps of a fictitious city. One half of the participants were instructed to learn their map as preparation to draw it from memory as precisely as possible (PrepDraw), which should stimulate the creation of an elaborated surface representation. The other half were instructed to learn the map as preparation for finding the shortest traffic connection from various locations to other locations (PrepConnect), which should stimulate the construction of a task-oriented deep-structure representation (mental model). Within both experimental groups, one-third of the participants received the map without prompts. Another third received the map with survey prompts (stimulating processing of what is where), and the final third received the map with connect prompts (stimulating processing of how train stations are connected). In the following test phase, participants received a recognition task, a recall task, and an inference task. For recognition and recall, two surface structure scores (extent, accuracy) and two deep structure scores (extent, accuracy) were calculated. The inference task served also to indicate deep structure accuracy. The PrepDraw group outperformed the PrepConnect group in terms of surface structure related variables, whereas the PrepConnect group outperformed the PrepDraw group in terms of deep structure-related variables. Map processing was not enhanced by prompts aligned with the instruction, but non-aligned prompts tended to interfere with learning.

Triggered kV Imaging During Spine SBRT for Intrafraction Motion Management

Purpose: To monitor intrafraction motion during spine stereotactic body radiotherapy(SBRT) treatment delivery with readily available technology, we implemented triggered kV imaging using the on-board imager(OBI) of a modern medical linear accelerator with an advanced imaging package. Methods: Triggered kV imaging for intrafraction motion management was tested with an anthropomorphic phantom and simulated spine SBRT treatments to the thoracic and lumbar spine. The vertebral bodies and spinous processes were contoured as the image guided radiotherapy(IGRT) structures specific to this technique. Upon each triggered kV image acquisition, 2D projections of the IGRT structures were automatically calculated and updated at arbitrary angles for display on the kV images. Various shifts/rotations were introduced in x, y, z, pitch, and yaw. Gantry-angle-based triggering was set to acquire kV images every 45°. A group of physicists/physicians(n = 10) participated in a survey to evaluate clinical efficiency and accuracy of clinical decisions on images containing various phantom shifts. This method was implemented clinically for treatment of 42 patients(94 fractions) with 15 second time-based triggering. Result: Phantom images revealed that IGRT structure accuracy and therefore utility of projected contours during triggered imaging improved with smaller CT slice thickness. Contouring vertebra superior and inferior to the treatment site was necessary to detect clinically relevant phantom rotation. From the survey, detectability was proportional to the shift size in all shift directions and inversely related to the CT slice thickness. Clinical implementation helped evaluate robustness of patient immobilization. Based on visual inspection of projected IGRT contours on planar kV images, appreciable intrafraction motion was detected in eleven fractions(11.7%). Discussion: Feasibility of triggered imaging for spine SBRT intrafraction motion management has been demonstrated in phantom experiments and implementation for patient treatments. This technique allows efficient, non-invasive monitoring of patient position using the OBI and patient anatomy as a direct visual guide.

Sensitivity analysis of truck structure based on CA algorithm

With the continuous improvement of modern CAE technology, structural reanalysis algorithm has gradually come into people’s vision and developed rapidly. The structure reanalysis algorithm introduced in this paper is an accelerated calculation method. The core idea of this algorithm is to avoid the complete analytical calculations after the structure modification, and reduce the calculation scale, save the calculation time, improve the efficiency of CAE simulation effectively on the premise of meeting the requirements of structure accuracy. The objective of this paper that is based on the initial third-order modal information of the truck structure is to control the overall quality of the structure. And it has important guiding significance for practical production. In this paper, different design variables are set in combination with the structural reanalysis algorithm. While the parameters of design variables are modified, sensitivity information analysis and Taylor expansion theorem are used to verify the feasibility and accuracy of the structural reanalysis method in optimal calculation

ProGen: Language Modeling for Protein Generation

Generative modeling for protein engineering is key to solving fundamental problems in synthetic biology, medicine, and material science. We pose protein engineering as an unsupervised sequence generation problem in order to leverage the exponentially growing set of proteins that lack costly, structural annotations. We train a 1.2B-parameter language model, ProGen, on ∼280M protein sequences conditioned on taxonomic and keyword tags such as molecular function and cellular component. This provides ProGen with an unprecedented range of evolutionary sequence diversity and allows it to generate with fine-grained control as demonstrated by metrics based on primary sequence similarity, secondary structure accuracy, and conformational energy.

Multi-scale structural analysis of proteins by deep semantic segmentation

Motivation Recent advances in computational methods have facilitated large-scale sampling of protein structures, leading to breakthroughs in protein structural prediction and enabling de novo protein design. Establishing methods to identify candidate structures that can lead to native folds or designable structures remains a challenge, since few existing metrics capture high-level structural features such as architectures, folds and conformity to conserved structural motifs. Convolutional Neural Networks (CNNs) have been successfully used in semantic segmentation—a subfield of image classification in which a class label is predicted for every pixel. Here, we apply semantic segmentation to protein structures as a novel strategy for fold identification and structure quality assessment. Results We train a CNN that assigns each residue in a multi-domain protein to one of 38 architecture classes designated by the CATH database. Our model achieves a high per-residue accuracy of 90.8% on the test set (95.0% average per-class accuracy; 87.8% average per-structure accuracy). We demonstrate that individual class probabilities can be used as a metric that indicates the degree to which a randomly generated structure assumes a specific fold, as well as a metric that highlights non-conformative regions of a protein belonging to a known class. These capabilities yield a powerful tool for guiding structural sampling for both structural prediction and design. Availability and implementation The trained classifier network, parser network, and entropy calculation scripts are available for download at https://git.io/fp6bd, with detailed usage instructions provided at the download page. A step-by-step tutorial for setup is provided at https://goo.gl/e8GB2S. All Rosetta commands, RosettaRemodel blueprints, and predictions for all datasets used in the study are available in the Supplementary Information. Supplementary information Supplementary data are available at Bioinformatics online.

Ontology Development for Medical and Health Science Domain in the Quran

Ontology is known as a knowledge representation and acts as a sharing platform for common ideas within a similar domain. It has a tree structure to ease the information presentation to users. Nowadays, it is very important to have a consistent and systematic way of presenting and retrieving different sources of knowledge such as the Quran and Hadith. Since there is so much useful information that can be retrieved from the Quran, especially for the Medical and Health Science domain, this paper presents the development of ontology for the Medical and Health Science domain in the Quran by adopting the Ontology 101 approach. These include the scope and domain determination, competency question formulation, ontology construction, and ontology evaluation. The proposed ontology in this paper has successfully retrieved the correct answers for Medical and Health Science using related queries via SPARQL-query and has been evaluated by the domain experts. Furthermore, the ontology structure accuracy has also been verified using reasoner, where it detected inconstancy during ontology development. For future work, this research paper can be used as a reference and basis to answer user queries, data integration with other applications or this ontology can be further expanded.