RESUMO
PrankWeb is an online resource providing an interface to P2Rank, a state-of-the-art method for ligand binding site prediction. P2Rank is a template-free machine learning method based on the prediction of local chemical neighborhood ligandability centered on points placed on a solvent-accessible protein surface. Points with a high ligandability score are then clustered to form the resulting ligand binding sites. In addition, PrankWeb provides a web interface enabling users to easily carry out the prediction and visually inspect the predicted binding sites via an integrated sequence-structure view. Moreover, PrankWeb can determine sequence conservation for the input molecule and use this in both the prediction and result visualization steps. Alongside its online visualization options, PrankWeb also offers the possibility of exporting the results as a PyMOL script for offline visualization. The web frontend communicates with the server side via a REST API. In high-throughput scenarios, therefore, users can utilize the server API directly, bypassing the need for a web-based frontend or installation of the P2Rank application. PrankWeb is available at http://prankweb.cz/, while the web application source code and the P2Rank method can be accessed at https://github.com/jendelel/PrankWebApp and https://github.com/rdk/p2rank, respectively.
Assuntos
Aprendizado de Máquina , Proteínas/química , Software , Sequência de Aminoácidos , Benchmarking , Sítios de Ligação , Conjuntos de Dados como Assunto , Humanos , Internet , Ligantes , Ligação Proteica , Conformação Proteica em alfa-Hélice , Conformação Proteica em Folha beta , Domínios e Motivos de Interação entre Proteínas , Proteínas/metabolismo , TermodinâmicaRESUMO
PURPOSE: To train a CycleGAN on downscaled versions of mammographic data to artificially inject or remove suspicious features, and to determine whether these AI-mediated attacks can be detected by radiologists. MATERIAL AND METHODS: From two publicly available datasets, BCDR and INbreast, we selected 680 images with and without lesions as training data. An internal dataset (nâ¯=â¯302 cancers, nâ¯=â¯590 controls) served as test data. We ran two experiments (256â¯×â¯256 px and 512â¯×â¯408 px) and applied the trained model to the test data. Three radiologists read a set of images (modified and originals) and rated the presence of suspicious lesions on a scale from 1 to 5 and the likelihood of the image being manipulated. The readout was evaluated by multiple reader multiple case receiver operating characteristics (MRMC-ROC) analysis using the area under the curve (AUC). RESULTS: At the lower resolution, the overall performance was not affected by the CycleGAN modifications (AUC 0.70 vs. 0.76, pâ¯=â¯0.67). However, one radiologist exhibited lower detection of cancer (0.85 vs 0.63, pâ¯=â¯0.06). The radiologists could not discriminate between original and modified images (0.55, pâ¯=â¯0.45). At the higher resolution, all radiologists showed significantly lower detection rate of cancer in the modified images (0.80 vs. 0.37, pâ¯<â¯0.001), however, they were able to detect modified images due to better visibility of artifacts (0.94, pâ¯<â¯0.0001). CONCLUSION: Our proof-of-concept study shows that CycleGAN can implicitly learn suspicious features and artificially inject or remove them in existing images. The applicability of the method is currently limited by the small image size and introduction of artifacts.