A transformer architecture for retention time prediction in liquid chromatography mass spectrometry-based proteomics

Accurate retention time (RT) prediction is important for spectral library-based analysis in data-independent acquisition mass spectrometry-based proteomics. The deep learning approach has demonstrated superior performance over traditional machine learning methods for this purpose. The transformer architecture is a recent development in deep learning that delivers state-of-the-art performance in many fields such as natural language processing, computer vision, and biology. We assess the performance of the transformer architecture for RT prediction using datasets from five deep learning models Prosit, DeepDIA, AutoRT, DeepPhospho, and AlphaPeptDeep. The experimental results on holdout datasets and independent datasets exhibit state-of-the-art performance of the transformer architecture. The software and evaluation datasets are publicly available for future development in the field.     

Conformational analysis and clustering of short and medium size loops connecting regular secondary structures: a database for modeling and prediction.

Loops are regions of nonrepetitive conformation connecting regular secondary structures. We identified 2,024 loops of one to eight residues in length, with acceptable main-chain bond lengths and peptide bond angles, from a database of 223 protein and protein-domain structures. Each loop is characterized by its sequence, main-chain conformation, and relative disposition of its bounding secondary structures as described by the separation between the tips of their axes and the angle between them. Loops, grouped according to their length and type of their bounding secondary structures, were superposed and clustered into 161 conformational classes, corresponding to 63% of all loops. Of these, 109 (51% of the loops) were populated by at least four nonhomologous loops or four loops sharing a low sequence identity. Another 52 classes, including 12% of the loops, were populated by at least three loops of low sequence similarity from three or fewer nonhomologous groups. Loop class suprafamilies resulting from variations in the termini of secondary structures are discussed in this article. Most previously described loop conformations were found among the classes. New classes included a 2:4 type IV hairpin, a helix-capping loop, and a loop that mediates dinucleotide-binding. The relative disposition of bounding secondary structures varies among loop classes, with some classes such as beta-hairpins being very restrictive. For each class, sequence preferences as key residues were identified; those most frequently at these conserved positions than in proteins were Gly, Asp, Pro, Phe, and Cys. Most of these residues are involved in stabilizing loop conformation, often through a positive phi conformation or secondary structure capping. Identification of helix-capping residues and beta-breakers among the highly conserved positions supported our decision to group loops according to their bounding secondary structures. Several of the identified loop classes were associated with specific functions, and all of the member loops had the same function; key residues were conserved for this purpose, as is the case for the parvalbumin-like calcium-binding loops. A significant number, but not all, of the member loops of other loop classes had the same function, as is the case for the helix-turn-helix DNA-binding loops. This article provides a systematic and coherent conformational classification of loops, covering a broad range of lengths and all four combinations of bounding secondary structure types, and supplies a useful basis for modelling of loop conformations where the bounding secondary structures are known or reliably predicted.     

Functional analogs for the reduction of certain nitrogenase substrates. Are multiple sites within the Fe/Mo/S active center involved in the 6e– reduction of N2?

 Reactivity studies of clusters that contain the MFe₃S₄ cores (M = Mo, V) with catecholate, multicarboxylate (or DMF) ligands coordinated to the Mo (or V) atoms, and Cl ligands coordinated to the Fe atoms have been carried out. These studies show the M/Fe/S single cubane clusters to be effective catalysts in the reduction of nitrogenase substrates such as hydrazine, acetylene and protons to give ammonia, ethylene and dihydrogen respectively. The same molecules do not activate or catalyze the reduction of dinitrogen. The results indicate that the observed catalyses are occurring at the Mo (V) sites by a process that, in the case of hydrazine, involves substrate protonation prior to reduction. The facile catalytic reduction of hydrazine by clusters that contain coordinatively saturated polycarboxylate-bound Mo atoms is rationalized in terms of a possible protonation/proton delivery function of the coordinated polycarboxylate ligands. The reactivity characteristics of the M/Fe/S clusters (structurally quite similar to the nitrogenase cofactor) have led to the suggestion that the Mo (V) atoms may be involved in the reduction of hydrazine in the later stages of dinitrogen reduction. Received and accepted: 21 August 1996     

Factors controlling the growth of field populations of Hydrodictyon reticulatum in New Zealand

Since Hydrodictyon reticulatum was introduced to New Zealand it has spread rapidly and produced persistent annual nuisance growths in areas where nuisance algal had not occurred previously. Field bioassays were conducted at 10 sites between August 1993 and February 1995 to evaluate the seasonal growth patterns and the factors controlling growth under natural conditions. H. reticulatum exhibited a strong seasonal growth pattern with growth rates up to 0.33 doublings d^-1 from August to March, are duction in growth rate in April and little or no growth from May to July. The H. reticulatum present in New Zealand has are latively low requirement for dissolved inorganic nitrogen (DIN) in comparison with other nuisance species, with its growth rate being saturated at 200 mg m^-3. This and the high affinity for DIN as shown by a K_s of 29 mg m^-3 have been key factors in the establishment of nuisance growths of H. reticulatum in New Zealand. This revised version was published online in August 2006 with corrections to the Cover Date.     

X-ray structure of recombinant horse L-chain apoferritin at 2.0 Å resolution: implications for stability and function

The X-ray structure of recombinant horse L-chain (rL) apoferritin, solved at 2.0?Å resolution with a final R factor of 17.9%, gives evidence that the residue at position 93 in the sequence is a proline and not a leucine, as found in earlier sequencing studies. The structure is isomorphous with other apoferritin structures, and we thus draw particular attention to those structural features which can be related to the stability and function of the protein. Analysis of hydrogen bonding and salt bridge interactions shows that dimers and tetramers are the most stable molecular entities within the protein shell: a result confirming earlier biophysical experiments. The stability of horse rL apoferritin to both dissociation into subunits at acidic pH values and to complete unfolding in guanidine chloride solutions is compared with that of other apoferritins. This emphasizes the role played by the salt bridge in the stability of this protein family. The horse rL apoferritin is significantly more resistant to denaturation than horse spleen ferritin, which in turn is more resistant than any human rH apoferritins, even those for which a salt bridge is restored. Finally, this structure determination not only establishes that a preformed pocket exists in L-chain apoferritin, at a site known to be able to bind porphyrin, but also underlines the particular function of a cluster of glutamic acids (E53, E56, E57 and E60) located at the entrance of this porphyrin-binding pocket.     

Sequence analysis of the AAA protein family.

The AAA protein family, a recently recognized group of Walker-type ATPases, has been subjected to an extensive sequence analysis. Multiple sequence alignments revealed the existence of a region of sequence similarity, the so-called AAA cassette. The borders of this cassette were localized and within it, three boxes of a high degree of conservation were identified. Two of these boxes could be assigned to substantial parts of the ATP binding site (namely, to Walker motifs A and B); the third may be a portion of the catalytic center. Phylogenetic trees were calculated to obtain insights into the evolutionary history of the family. Subfamilies with varying degrees of intra-relatedness could be discriminated; these relationships are also supported by analysis of sequences outside the canonical AAA boxes: within the cassette are regions that are strongly conserved within each subfamily, whereas little or even no similarity between different subfamilies can be observed. These regions are well suited to define fingerprints for subfamilies. A secondary structure prediction utilizing all available sequence information was performed and the result was fitted to the general 3D structure of a Walker A/GTPase. The agreement was unexpectedly high and strongly supports the conclusion that the AAA family belongs to the Walker superfamily of A/GTPases.     

The membrane topology of the amino-terminal domain of the red cell calcium pump.

A systematic study of the membrane-associated regions in the plasma membrane Ca2+ pump of erythrocytes has been performed by hydrophobic photolabeling. Purified Ca2+ pump was labeled with 3-(trifluoromethyl)-3-(m-[125I]iodophenyl)-diazirine ([125I]TID), a generic photoactivatable hydrophobic probe. These results were compared with the enzyme labeled with a strictly membrane-bound probe, [3H]bis-phosphatidylethanolamine (trifluoromethyl) phenyldiazirine. A significant light-dependent labeling of an M(r) 135,000-140,000 peptide, corresponding to the full Ca2+ pump, was observed with both probes. After proteolysis of the pump labeled with each probe and isolation of fragments by SDS-PAGE, a common pattern of labeled peptides was observed. Similarly, labeling of the Ca2+ pump with [125I]TID, either in isolated red blood cell membranes or after the enzyme was purified, yields a similar pattern of labeled peptides. Taken together, these results validate the use of either probe to study the lipid interface of the membrane-embedded region of this protein, and sustain the notion that the conformation of the pump is maintained throughout the procedures of solubilization, affinity purification, and reconstitution into proteoliposomes. In this work, we put special emphasis on a detailed analysis of the N-terminal domain of the Ca2+ pump. A labeled peptide of M(r) 40,000 belonging to this region was purified and further digested with V8 protease. The specific incorporation of [125I]TID to proteolytic fragments pertaining to the amino-terminal region indicates the existence of two transmembrane stretches in this domain. A theoretical analysis based on the amino acid sequence 1-322 predicts two segments with high probability of membrane insertion, in agreement with the experimental data. Each segment shows a periodicity pattern of hydrophobicity and variability compatible with alpha-helical structure. These results strongly suggest the existence of a transmembrane helical hairpin motif near the N-terminus of the Ca2+ pump.     

Three-dimensional structure prediction of the NAD binding site of proton-pumping transhydrogenase from Escherichia coli

A three-dimensional structure of the NAD site of Escerichia coli transhydrogenase has been predicted. The model is based on analysis of conserved residues among the transhydrogenases from five different sources, homologies with enzymes using NAD as cofactors or substrates, hydrophilicity profiles, and secondary structure predictions. The present model supports the hypothesis that there is one binding site, located relatively close to the N-terminus of the α-subunit. The proposed structure spans residues α145 to α287, and it includes five β-strands and five α-helices oriented in a typical open twisted α/β conformation. The amino acid sequence following the GXGXXG dinucleotide binding consensus sequence (residues α172 to α177) correlates exactly to a typical fingerprint region for ADP binding βαβ folds in dinucleotide binding enzymes. In the model, aspartic acid α195 forms hydrogen bonds to one or both hydroxyl groups on the adenosine ribose sugar moiety. Threonine α196 and alanine α256, located at the end of βB and βD, respectively, create a hydrophobic sandwich with the adenine part of NAD buried inside. The nicotinamide part is located in a hydrophobic cleft between αA and βE. Mutagenesis work has been carried out in order to test the predicted model and to determine whether residues within this domain are important for proton pumping directly. All data support the predicted structure, and no residue crucial for proton pumping Was detected. Since no three-dimensional structure of transhydrogenase has been solved, a well based tertiary structure prediction is of great value for further experimental design in trying to elucidate the mechanism of the energy-linked proton pump. © 1995 Wiley-Liss, Inc.     

Knowledge-based modeling of the D-lactate dehydrogenase three-dimensional structure

A three-dimensional structure of the NAD-dependent D -lactate dehydrogenase of Lactobacillus bulgaricus is modeled using the structure of the formate dehydrogenase of Pseudomonas sp. as template. Both sequences share only 22% of identical residues. Regions for knowledge-based modeling are defined from the structurally conserved regions predicted by multiple alignment of a set of related protein sequences with low homology. The model of the D -LDH subunit shows, as for the formate dehydrogenase, an α/β structure, with a catalytic domain and a coenzyme binding domain. It points out the catalytic histidine (His-296) and supports the hypothetical catalytic mechanism. It also suggests that the other residues involved in the active site are Arg-235, possibly involved in the binding of the carboxyl group of the pyruvate, and Phe-299, a candidate for stabilizing the methyl group of the substrate. © 1995 Wiley-Liss, Inc.