CAID prediction portal: a comprehensive service for predicting intrinsic disorder and binding regions in proteins
Alessio Del Conte, Adel Bouhraoua, Mahta Mehdiabadi, Damiano Clementel, Alexander Miguel Monzón, CAID predictors, Alex S. Holehouse, Daniel Griffith, Ryan J. Emenecker, Ashwini Patil, Ronesh Sharma, Tatsuhiko Tsunoda, Alok Sharma, Yi Jun Tang, Bin Liu, Claudio Mirabello, Björn Wallner, Burkhard Rost, Dagmar Ilzhöfer, Maria Littmann, Michael Heinzinger, Lea I M Krautheimer, Michael Bernhofer, Liam J. McGuffin, Isabelle Callebaut, Tristan Bitard Feildel, Jian Liu, Jianlin Cheng, Zhiye Guo, Jinbo Xu, Sheng Wang, Nawar Malhis, Jörg Gsponer, Chol-Song Kim, Kun‐Sop Han, Myong-Chol Ma, Lukasz Kurgan, Sina Ghadermarzi, Akila Katuwawala, Bi Zhao, Zhenling Peng, Zhonghua Wu, Gang Hu, Kui Wang, Md Tamjidul Hoque, Md Wasi Ul Kabir, Michele Vendruscolo, Pietro Sormanni, Min Li, Fuhao Zhang, Pengzhen Jia, Yida Wang, Michail Yu. Lobanov, Oxana V. Galzitskaya, Wim Vranken, Adrián Díaz, Thomas Litfin, Yaoqi Zhou, Jack Hanson, Kuldip K. Paliwal, Zsuzsanna Dosztányi, Gábor Erdős, Silvio C. E. Tosatto, Damiano Piovesan
Abstract
Intrinsic disorder (ID) in proteins is well-established in structural biology, with increasing evidence for its involvement in essential biological processes. As measuring dynamic ID behavior experimentally on a large scale remains difficult, scores of published ID predictors have tried to fill this gap. Unfortunately, their heterogeneity makes it difficult to compare performance, confounding biologists wanting to make an informed choice. To address this issue, the Critical Assessment of protein Intrinsic Disorder (CAID) benchmarks predictors for ID and binding regions as a community blind-test in a standardized computing environment. Here we present the CAID Prediction Portal, a web server executing all CAID methods on user-defined sequences. The server generates standardized output and facilitates comparison between methods, producing a consensus prediction highlighting high-confidence ID regions. The website contains extensive documentation explaining the meaning of different CAID statistics and providing a brief description of all methods. Predictor output is visualized in an interactive feature viewer and made available for download in a single table, with the option to recover previous sessions via a private dashboard. The CAID Prediction Portal is a valuable resource for researchers interested in studying ID in proteins. The server is available at the URL: https://caid.idpcentral.org.