Improved prediction of protein secondary structure by use of sequence profiles and neural networks.

By Burkhard Rost, Christopher Sander. 1993

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Modeling aspects of the language of life through transfer-learning protein sequences (2019, methodology): “The second layer (struc-ture-to-structure [2, 3]) created the final predictions by applying again a CNN (w = 7) over the output of the first layer.”
CPHmodels-3.0—remote homology modeling using structure-guided sequence profiles (2010, background): “Sequence profiles have a broad application in the field of bioinformatics prediction algorithms dating back to the pioneering work by Rost and Sander (1).”
Computational prediction of secondary and supersecondary structures. (2013, methodology): “Using the multiple sequence alignment profile rather than the primary sequence has led to a large improvement by 10% accuracy in the secondary structure prediction (91).”
Artificial neural networks for computer-based molecular design. (1998, methodology): “A signi®cant contribution was done by Rost and Sander who introduced the multiple alignment of training sequences for a better overall accuracy and a balanced training (Rost and Sander, 1993a,b).”
Sixty-five years of the long march in protein secondary structure prediction: the final stretch? (2016, background): “The third generation of techniques is characterized by using evolutionary information derived from alignment of multiple homologous sequences [54, 75].”
Best α‐helical transmembrane protein topology predictions are achieved using hidden Markov models and evolutionary information (2004, background): “For instance, this is known to improve prediction performance of secondary structure prediction for globular proteins by several percentage units (Rost and Sander 1993).”
Local Protein Structures (2007, methodology): “This prediction combined a secondary structure prediction done by PHD [332, 333] and an adequacy score of the 82 clusters.”
Prediction of protein stability changes for single‐site mutations using support vector machines (2005, methodology): “This approach has been applied successfully to the prediction of other protein structural features, such as secondary structure and solvent accessibility [40, 41, 42, 43].”
Modeling the language of life – Deep Learning Protein Sequences (2019, methodology): “The second layer (structure-to-structure [1, 2]) created 33 the final predictions by applying again a CNN (w=7) over the output of the first layer.”
ProtTrans: Towards Cracking the Language of Life’s Code Through Self-Supervised Learning (2021, methodology): “Top prediction methods in computational biology [15], [16], [17], [18], [19], [20] combine machine learning (ML) and evolutionary information (EI), ﬁrst established as the winning strategy to predict protein secondary structure [21], [22] in two steps.”
Neural networks for secondary structure and structural class predictions (1995, methodology): “Compared with results obtained by classifying proteins directly using secondary structure predictions (Rost & Sander, 1993b), the 4-output network was comparable in classifying all-a! proteins (57% versus 58”
Exploiting the past and the future in protein secondary structure prediction (1999, unknown): “Building upon the work of Qian and Sejnowski (1988), for a long time the best SS prediction performance has been achieved by the PHD scheme (Rost and Sander, 1993, 1994).”
Transmembrane helices predicted at 95% accuracy (1995, methodology): “The principles of neural networks for secondary structure prediction (Fariselli et al., 1993, Rost & Sander, 1993a) and of coding multiple sequence information (Rost & Sander, 1993b, Rost & Sander, 1994a, Rost & Sander, 1994b) are described in detail elsewhere.”
Improving prediction of protein secondary structure, backbone angles, solvent accessibility and contact numbers by using predicted contact maps and an ensemble of recurrent and residual convolutional neural networks (2018, background): “The accuracy of secondary structure prediction has risen from 70% (Rost and Sander, 1993) to the latest 85% (Fang et al.”