EPSILON-CP: using deep learning to combine information from multiple sources for protein contact prediction.

BACKGROUND: Accurately predicted contacts allow to compute the 3D structure of a protein. Since the solution space of native residue-residue contact pairs is very large, it is necessary to leverage information to identify relevant regions of the solution space, i.e. correct contacts. Every additional source of information can contribute to narrowing down candidate regions. Therefore, recent methods combined evolutionary and sequence-based information as well as evolutionary and physicochemical information. We develop a new contact predictor (EPSILON-CP) that goes beyond current methods by combining evolutionary, physicochemical, and sequence-based information. The problems resulting from the increased dimensionality and complexity of the learning problem are combated with a careful feature analysis, which results in a drastically reduced feature set. The different information sources are combined using deep neural networks. RESULTS: On 21 hard CASP11 FM targets, EPSILON-CP achieves a mean precision of 35.7% for top- L/10 predicted long-range contacts, which is 11% better than the CASP11 winning version of MetaPSICOV. The improvement on 1.5L is 17%. Furthermore, in this study we find that the amino acid composition, a commonly used feature, is rendered ineffective in the context of meta approaches. The size of the refined feature set decreased by 75%, enabling a significant increase in training data for machine learning, contributing significantly to the observed improvements. CONCLUSIONS: Exploiting as much and diverse information as possible is key to accurate contact prediction. Simply merging the information introduces new challenges. Our study suggests that critical feature analysis can improve the performance of contact prediction methods that combine multiple information sources. EPSILON-CP is available as a webservice: http://compbio.robotics.tu-berlin.de/epsilon/.

Protein structure prediction with in-cell photo-crosslinking mass spectrometry and deep learning.

While AlphaFold2 can predict accurate protein structures from the primary sequence, challenges remain for proteins that undergo conformational changes or for which few homologous sequences are known. Here we introduce AlphaLink, a modified version of the AlphaFold2 algorithm that incorporates experimental distance restraint information into its network architecture. By employing sparse experimental contacts as anchor points, AlphaLink improves on the performance of AlphaFold2 in predicting challenging targets. We confirm this experimentally by using the noncanonical amino acid photo-leucine to obtain information on residue-residue contacts inside cells by crosslinking mass spectrometry. The program can predict distinct conformations of proteins on the basis of the distance restraints provided, demonstrating the value of experimental data in driving protein structure prediction. The noise-tolerant framework for integrating data in protein structure prediction presented here opens a path to accurate characterization of protein structures from in-cell data.

Blind testing cross-linking/mass spectrometry under the auspices of the 11th critical assessment of methods of protein structure prediction (CASP11).

Determining the structure of a protein by any method requires various contributions from experimental and computational sides. In a recent study, high-density cross-linking/mass spectrometry (HD-CLMS) data in combination with ab initio structure prediction determined the structure of human serum albumin (HSA) domains, with an RMSD to X-ray structure of up to 2.5 Å, or 3.4 Å in the context of blood serum. This paper reports the blind test on the readiness of this technology through the help of Critical Assessment of protein Structure Prediction (CASP). We identified between 201-381 unique residue pairs at an estimated 5% FDR (at link level albeit with missing site assignment precision evaluation), for four target proteins. HD-CLMS proved reliable once crystal structures were released. However, improvements in structure prediction using cross-link data were slight. We identified two reasons for this. Spread of cross-links along the protein sequence and the tightness of the spatial constraints must be improved. However, for the selected targets even ideal contact data derived from crystal structures did not allow modellers to arrive at the observed structure. Consequently, the progress of HD-CLMS in conjunction with computational modeling methods as a structure determination method, depends on advances on both arms of this hybrid approach.