Convolutional ProteinUnetLM competitive with long short-term memory-based protein secondary structure predictors.

The protein secondary structure (SS) prediction plays an important role in the characterization of general protein structure and function. In recent years, a new generation of algorithms for SS prediction based on embeddings from protein language models (pLMs) is emerging. These algorithms reach state-of-the-art accuracy without the need for time-consuming multiple sequence alignment (MSA) calculations. Long short-term memory (LSTM)-based SPOT-1D-LM and NetSurfP-3.0 are the latest examples of such predictors. We present the ProteinUnetLM model using a convolutional Attention U-Net architecture that provides prediction quality and inference times at least as good as the best LSTM-based models for 8-class SS prediction (SS8). Additionally, we address the issue of the heavily imbalanced nature of the SS8 problem by extending the loss function with the Matthews correlation coefficient, and by proper assessment using previously introduced adjusted geometric mean (AGM) metric. ProteinUnetLM achieved better AGM and sequence overlap score than LSTM-based predictors, especially for the rare structures 310-helix (G), beta-bridge (B), and high curvature loop (S). It is also competitive on challenging datasets without homologs, free-modeling targets, and chameleon sequences. Moreover, ProteinUnetLM outperformed its previous MSA-based version ProteinUnet2, and provided better AGM than AlphaFold2 for 1/3 of proteins from the CASP14 dataset, proving its potential for making a significant step forward in the domain. To facilitate the usage of our solution by protein scientists, we provide an easy-to-use web interface under https://biolib.com/SUT/ProteinUnetLM/.

New insights into disordered proteins and regions according to the FOD-M model.

A collection of intrinsically disordered proteins (IDPs) having regions with the status of intrinsically disordered (IDR) according to the Disprot database was analyzed from the point of view of the structure of hydrophobic core in the structural unit (chain / domain). The analysis includes all the Homo Sapiens as well as Mus Musculus proteins present in the DisProt database for which the structure is available. In the analysis, the fuzzy oil drop modified model (FOD-M) was used, taking into account the external force field, modified by the presence of other factors apart from polar water, influencing protein structuring. The paper presents an alternative to secondary-structure-based classification of intrinsically disordered regions (IDR). The basis of our classification is the ordering of hydrophobic core as calculated by the FOD-M model resulting in FOD-ordered or FOD-unordered IDRs.

Connexins and Pannexins-Similarities and Differences According to the FOD-M Model.

Connexins and pannexins are the transmembrane proteins of highly distinguished biological activity in the form of transport of molecules and electrical signals. Their common role is to connect the external environment with the cytoplasm of the cell, while connexin is also able to link two cells together allowing the transport from one to another. The analysis presented here aims to identify the similarities and differences between connexin and pannexin. As a comparative criterion, the hydrophobicity distribution in the structure of the discussed proteins was used. The comparative analysis is carried out with the use of a mathematical model, the FOD-M model (fuzzy oil drop model in its Modified version) expressing the specificity of the membrane's external field, which in the case of the discussed proteins is significantly different from the external field for globular proteins in the polar environment of water. The characteristics of the external force field influence the structure of protein allowing the activity in a different environment.

Changepoint Detection in Noisy Data Using a Novel Residuals Permutation-Based Method (RESPERM): Benchmarking and Application to Single Trial ERPs.

An important problem in many fields dealing with noisy time series, such as psychophysiological single trial data during learning or monitoring treatment effects over time, is detecting a change in the model underlying a time series. Here, we present a new method for detecting a single changepoint in a linear time series regression model, termed residuals permutation-based method (RESPERM). The optimal changepoint in RESPERM maximizes Cohen's effect size with the parameters estimated by the permutation of residuals in a linear model. RESPERM was compared with the SEGMENTED method, a well-established and recommended method for detecting changepoints, using extensive simulated data sets, varying the amount and distribution characteristics of noise and the location of the change point. In time series with medium to large amounts of noise, the variance of the detected changepoint was consistently smaller for RESPERM than SEGMENTED. Finally, both methods were applied to a sample dataset of single trial amplitudes of the N250 ERP component during face learning. In conclusion, RESPERM appears to be well suited for changepoint detection especially in noisy data, making it the method of choice in neuroscience, medicine and many other fields.

On the need to introduce environmental characteristics in ab initio protein structure prediction using a coarse-grained UNRES force field.

During the protein folding process in computer simulations involving the use of a United RESidue (UNRES) force field, an additional module was introduced to represent directly the presence of a polar solvent in water form. This module implements the fuzzy oil drop model (FOD) where the 3D Gauss function expresses the presence of a polar environment which directs the polypeptide chain folding process towards the generation of a centric hydrophobic core. Sample test polypeptide chains of 8 proteins with chain lengths ranging from 37 to 75 aa were simulated in silico using the UNRES (U) package with an implicit solvent model and a built-in module expressing the FOD model (UNRES-FOD-UNRES (U + F) interleaved simulation). The protein structure obtained by both *** simulation schemes, i.e., accordingly***U and U + F, for all the analyzed protein models shows the presence of a hydrophobic core including where it is absent in the native structure. The proposed FOD-M model (M-modified) explaining the source of this phenomenon reveals the need to modify the external field expressing the role of a folding environment. The modification takes into account the influence of other than polar factors present in the folding environment.

On the Dependence of Prion and Amyloid Structure on the Folding Environment.

Currently available analyses of amyloid proteins reveal the necessity of the existence of radical structural changes in amyloid transformation processes. The analysis carried out in this paper based on the model called fuzzy oil drop (FOD) and its modified form (FOD-M) allows quantifying the role of the environment, particularly including the aquatic environment. The starting point and basis for the present presentation is the statement about the presence of two fundamentally different methods of organizing polypeptides into ordered conformations-globular proteins and amyloids. The present study shows the source of the differences between these two paths resulting from the specificity of the external force field coming from the environment, including the aquatic and hydrophobic one. The water environment expressed in the fuzzy oil drop model using the 3D Gauss function directs the folding process towards the construction of a micelle-like system with a hydrophobic core in the central part and the exposure of polarity on the surface. The hydrophobicity distribution of membrane proteins has the opposite characteristic: Exposure of hydrophobicity at the surface of the membrane protein with an often polar center (as in the case of ion channels) is expected. The structure of most proteins is influenced by a more or less modified force field generated by water through the appropriate presence of a non-polar (membrane-like) environment. The determination of the proportion of a factor different from polar water enables the assessment of the protein status by indicating factors favoring the structure it represents.

In Silico Modeling of the Influence of Environment on Amyloid Folding Using FOD-M Model.

The role of the environment in amyloid formation based on the fuzzy oil drop model (FOD) is discussed here. This model assumes that the hydrophobicity distribution within a globular protein is consistent with a 3D Gaussian (3DG) distribution. Such a distribution is interpreted as the idealized effect of the presence of a polar solvent-water. A chain with a sequence of amino acids (which are bipolar molecules) determined by evolution recreates a micelle-like structure with varying accuracy. The membrane, which is a specific environment with opposite characteristics to the polar aquatic environment, directs the hydrophobic residues towards the surface. The modification of the FOD model to the FOD-M form takes into account the specificity of the cell membrane. It consists in "inverting" the 3DG distribution (complementing the Gaussian distribution), which expresses the exposure of hydrophobic residues on the surface. It turns out that the influence of the environment for any protein (soluble or membrane-anchored) is the result of a consensus factor expressing the participation of the polar environment and the "inverted" environment. The ratio between the proportion of the aqueous and the "reversed" environment turns out to be a characteristic property of a given protein, including amyloid protein in particular. The structure of amyloid proteins has been characterized in the context of prion, intrinsically disordered, and other non-complexing proteins to cover a wider spectrum of molecules with the given characteristics based on the FOD-M model.

The Functional Significance of Hydrophobic Residue Distribution in Bacterial Beta-Barrel Transmembrane Proteins.

ß-barrel membrane proteins have several important biological functions, including transporting water and solutes across the membrane. They are active in the highly hydrophobic environment of the lipid membrane, as opposed to soluble proteins, which function in a more polar, aqueous environment. Globular soluble proteins typically have a hydrophobic core and a polar surface that interacts favorably with water. In the fuzzy oil drop (FOD) model, this distribution is represented by the 3D Gauss function (3DG). In contrast, membrane proteins expose hydrophobic residues on the surface, and, in the case of ion channels, the polar residues face inwards towards a central pore. The distribution of hydrophobic residues in membrane proteins can be characterized by means of 1-3DG, a complementary 3D Gauss function. Such an analysis was carried out on the transmembrane proteins of bacteria, which, despite the considerable similarities of their super-secondary structure (ß-barrel), have highly differentiated properties in terms of stabilization based on hydrophobic interactions. The biological activity and substrate specificity of these proteins are determined by the distribution of the polar and nonpolar amino acids. The present analysis allowed us to compare the ways in which the different proteins interact with antibiotics and helped us understand their relative importance in the development of the resistance mechanism. We showed that beta barrel membrane proteins with a hydrophobic core interact less strongly with the molecules they transport.

Solubility and Aggregation of Selected Proteins Interpreted on the Basis of Hydrophobicity Distribution.

Protein solubility is based on the compatibility of the specific protein surface with the polar aquatic environment. The exposure of polar residues to the protein surface promotes the protein's solubility in the polar environment. The aquatic environment also influences the folding process by favoring the centralization of hydrophobic residues with the simultaneous exposure to polar residues. The degree of compatibility of the residue distribution, with the model of the concentration of hydrophobic residues in the center of the molecule, with the simultaneous exposure of polar residues is determined by the sequence of amino acids in the chain. The fuzzy oil drop model enables the quantification of the degree of compatibility of the hydrophobicity distribution observed in the protein to a form fully consistent with the Gaussian 3D function, which expresses an idealized distribution that meets the preferences of the polar water environment. The varied degrees of compatibility of the distribution observed with the idealized one allow the prediction of preferences to interactions with molecules of different polarity, including water molecules in particular. This paper analyzes a set of proteins with different levels of hydrophobicity distribution in the context of the solubility of a given protein and the possibility of complex formation.

The N250 event-related potential as an index of face familiarity: a replication study.

The neural correlates of face individuation-the acquisition of memory representations for novel faces-have been studied only in coarse detail and disregarding individual differences between learners. In their seminal study, Tanaka et al. (Tanaka et al. 2006 J. Cogn. Neurosci. 18, 1488-1497. (doi:10.1162/jocn.2006.18.9.1488)) required the identification of a particular novel face across 70 trials and found that the N250 component in the EEG event-related potentials became more negative from the first to the second half of the experiment, where it reached a similar amplitude as a well-known face. We were unable to directly replicate this finding in our study when we used the original split of trials. However, when we applied a different split of trials we observed very similar changes in N250 amplitude. We conclude that the N250 component is indeed sensitive to the build-up of a robust representation of a face in memory; the time course of this process appears to vary as a function of variables that may be determined in future research.

Model of Environmental Membrane Field for Transmembrane Proteins.

The water environment determines the activity of biological processes. The role of such an environment interpreted in the form of an external field expressed by the 3D Gaussian distribution in the fuzzy oil drop model directs the folding process towards the generation of a centrally located hydrophobic core with the simultaneous exposure of polar residues on the surface. In addition to proteins soluble in the water environment, there is a significant group of membrane proteins that act as receptors or channels, including ion channels in particular. The change of the polar (water) environment into a highly hydrophobic (membrane) environment is quite radical, resulting in a different hydrophobicity distribution within the membrane protein. Modification of the notation of the force field expressing the presence of the hydrophobic environment has been proposed in this work. A modified fuzzy oil drop model with its adaptation to membrane proteins was used to interpret the structure of membrane proteins-mechanosensitive channel. The modified model was also used to describe the so-called negative cases-i.e., for water-soluble proteins with a clear distribution consistent with the fuzzy oil drop model.

Divergence Entropy-Based Evaluation of Hydrophobic Core in Aggressive and Resistant Forms of Transthyretin.

The two forms of transthyretin differing slightly in the tertiary structure, despite the presence of five mutations, show radically different properties in terms of susceptibility to the amyloid transformation process. These two forms of transthyretin are the object of analysis. The search for the sources of these differences was carried out by means of a comparative analysis of the structure of these molecules in their native and early intermediate stage forms in the folding process. The criterion for assessing the degree of similarity and differences is the status of the hydrophobic core. The comparison of the level of arrangement of the hydrophobic core and its initial stages is possible thanks to the application of divergence entropy for the early intermediate stage and for the final forms. It was shown that the minimal differences observed in the structure of the hydrophobic core of the forms available in PDB, turned out to be significantly different in the early stage (ES) structure in folding process. The determined values of divergence entropy for both ES forms indicate the presence of the seed of hydrophobic core only in the form resistant to amyloid transformation. In the form of aggressively undergoing amyloid transformation, the structure lacking such a seed is revealed, being a stretched one with a high content of ß-type structure. In the discussed case, the active presence of water in the structural transformation of proteins expressed in the fuzzy oil drop model (FOD) is of decisive importance for the generation of the final protein structure. It has been shown that the resistant form tends to generate a centric hydrophobic core with the possibility of creating a globular structure, i.e., a spherical micelle-like form. The aggressively transforming form reveals in the structure of its early intermediate, a tendency to form the ribbon-like micelle as observed in amyloid.

Dependence of Protein Structure on Environment: FOD Model Applied to Membrane Proteins.

The natural environment of proteins is the polar aquatic environment and the hydrophobic (amphipathic) environment of the membrane. The fuzzy oil drop model (FOD) used to characterize water-soluble proteins, as well as its modified version FOD-M, enables a mathematical description of the presence and influence of diverse environments on protein structure. The present work characterized the structures of membrane proteins, including those that act as channels, and a water-soluble protein for contrast. The purpose of the analysis was to verify the possibility that an external force field can be used in the simulation of the protein-folding process, taking into account the diverse nature of the environment that guarantees a structure showing biological activity.

Downhill, Ultrafast and Fast Folding Proteins Revised.

Research on the protein folding problem differentiates the protein folding process with respect to the duration of this process. The current structure encoded in sequence dogma seems to be clearly justified, especially in the case of proteins referred to as fast-folding, ultra-fast-folding or downhill. In the present work, an attempt to determine the characteristics of this group of proteins using fuzzy oil drop model is undertaken. According to the fuzzy oil drop model, a protein is a specific micelle composed of bi-polar molecules such as amino acids. Protein folding is regarded as a spherical micelle formation process. The presence of covalent peptide bonds between amino acids eliminates the possibility of free mutual arrangement of neighbors. An example would be the construction of co-micelles composed of more than one type of bipolar molecules. In the case of fast folding proteins, the amino acid sequence represents the optimal bipolarity system to generate a spherical micelle. In order to achieve the native form, it is enough to have an external force field provided by the water environment which directs the folding process towards the generation of a centric hydrophobic core. The influence of the external field can be expressed using the 3D Gaussian function which is a mathematical model of the folding process orientation towards the concentration of hydrophobic residues in the center with polar residues exposed on the surface. The set of proteins under study reveals a hydrophobicity distribution compatible with a 3D Gaussian distribution, taken as representing an idealized micelle-like distribution. The structure of the present hydrophobic core is also discussed in relation to the distribution of hydrophobic residues in a partially unfolded form.

The Structure of Amyloid Versus the Structure of Globular Proteins.

The issue of changing the structure of globular proteins into an amyloid form is in the focus of researchers' attention. Numerous experimental studies are carried out, and mathematical models to define the essence of amyloid transformation are sought. The present work focuses on the issue of the hydrophobic core structure in amyloids. The form of ordering the hydrophobic core in globular proteins is described by a 3D Gaussian distribution analog to the distribution of hydrophobicity in a spherical micelle. Amyloid fibril is a ribbon-like micelle made up of numerous individual chains, each representing a flat structure. The distribution of hydrophobicity within a single chain included in the fibril describes the 2D Gaussian distribution. Such a description expresses the location of polar residues on a circle with a center with a high level of hydrophobicity. The presence of this type of order in the amyloid forms available in Preotin Data Bank (PDB) (both in proto- and superfibrils) is demonstrated in the present work. In this system, it can be assumed that the amyloid transformation is a chain transition from 3D Gauss ordering to 2D Gauss ordering. This means changing the globular structure to a ribbon-like structure. This observation can provide a simple mathematical model for simulating the amyloid transformation of proteins.

Structure of the Hydrophobic Core Determines the 3D Protein Structure-Verification by Single Mutation Proteins.

Four de novo proteins differing in single mutation positions, with a chain length of 56 amino acids, represent diverse 3D structures: monomeric 3α and 4ß + α folds. The reason for this diversity is seen in the different structure of the hydrophobic core as a result of synergy leading to the generation of a system in which the polypeptide chain as a whole participates. On the basis of the fuzzy oil drop model, where the structure of the hydrophobic core is expressed by means of the hydrophobic distribution function in the form of a 3D Gaussian distribution, it has been shown that the composition of the hydrophobic core in these two structural forms is different. In addition, the use of a model to determine the structure of the early intermediate in the folding process allows to indicate differences in the polypeptide chain geometry, which, combined with the construction of a common hydrophobic nucleus as an effect of specific synergy, may indicate the reason for the diversity of the folding process of the polypeptide chain. The results indicate the need to take into account the presence of an external force field originating from the water environment and that its active impact on the formation of a hydrophobic core whose participation in the stabilization of the tertiary structure is fundamental.

Model of Early Stage Intermediate in Respect to Its Final Structure.

The model, describing a method of determining the structure of an early intermediate in the process of protein folding to analyze nonredundant PDB protein bases, allows determining the relationship between the sequence of tetrapeptides and their structural forms expressed by structural codes. The contingency table expressing such a relationship can be used to predict the structure of polypeptides by proposing a structural form with a precision limited to the structural code. However, by analyzing structural forms in native forms of proteins based on the fuzzy oil drop model, one can also determine the status of polypeptide chain fragments with respect to the assumptions of this model. Whether the probability distributions for both compliant and noncompliant forms were similar or whether the tetrapeptide sequences showed some differences at a level of a set of structural codes was investigated. The analysis presented here indicated that some sequences in both forms revealed differences in probability distributions expressed as a negative statistically significant correlation coefficient. This meant that the identified sections (tetrapeptides) took different forms against the fuzzy oil drop model. It may suggest that the information of the final status with respect to hydrophobic core formation is already carried by the structure of the early-stage intermediate.

Different Synergy in Amyloids and Biologically Active Forms of Proteins.

Protein structure is the result of the high synergy of all amino acids present in the protein. This synergy is the result of an overall strategy for adapting a specific protein structure. It is a compromise between two trends: The optimization of non-binding interactions and the directing of the folding process by an external force field, whose source is the water environment. The geometric parameters of the structural form of the polypeptide chain in the form of a local radius of curvature that is dependent on the orientation of adjacent peptide bond planes (result of the respective Phi and Psi rotation) allow for a comparative analysis of protein structures. Certain levels of their geometry are the criteria for comparison. In particular, they can be used to assess the differences between the structural form of biologically active proteins and their amyloid forms. On the other hand, the application of the fuzzy oil drop model allows the assessment of the role of amino acids in the construction of tertiary structure through their participation in the construction of a hydrophobic core. The combination of these two models-the geometric structure of the backbone and the determining of the participation in the construction of the tertiary structure that is applied for the comparative analysis of biologically active and amyloid forms-is presented.

Statistical dictionaries for hypothetical in silico model of the early-stage intermediate in protein folding.

The polypeptide chain folding process appears to be a multi-stage phenomenon. The scientific community has recently devoted much attention to early stages of this process, with numerous attempts at simulating them--either experimentally or in silico. This paper presents a comparative analysis of the predicted and observed results of folding simulations. The proposed technique, based on statistical dictionaries, yields a global accuracy of 57%--a marked improvement over older approaches (with an accuracy of approximately 46%).