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.     

Evolutionary link between glycogen phosphorylase and a DNA modifying enzyme.

We report here an unexpected similarity in three-dimensional structure between glucosyltransferases involved in very different biochemical pathways, with interesting evolutionary and functional implications. One is the DNA modifying enzyme beta-glucosyltransferase from bacteriophage T4, alias UDP-glucose:5-hydroxymethyl-cytosine beta-glucosyltransferase. The other is the metabolic enzyme glycogen phosphorylase, alias 1.4-alpha-D-glucan:orthophosphate alpha-glucosyltransferase. Structural alignment revealed that the entire structure of beta-glucosyltransferase is topographically equivalent to the catalytic core of the much larger glycogen phosphorylase. The match includes two domains in similar relative orientation and connecting helices, with a positional root-mean-square deviation of only 3.4 A for 256 C alpha atoms. An interdomain rotation seen in the R- to T-state transition of glycogen phosphorylase is similar to that observed in beta-glucosyltransferase on substrate binding. Although not a single functional residue is identical, there are striking similarities in the spatial arrangement and in the chemical nature of the substrates. The functional analogies are (beta-glucosyltransferase-glycogen phosphorylase): ribose ring of UDP-pyridoxal ring of pyridoxal phosphate co-enzyme; phosphates of UDP-phosphate of co-enzyme and reactive orthophosphate; glucose unit transferred to DNA-terminal glucose unit extracted from glycogen. We anticipate the discovery of additional structurally conserved members of the emerging glucosyltransferase superfamily derived from a common ancient evolutionary ancestor of the two enzymes.     

A Drosophila hsp70 gene contains long,antiparallel, coupled open reading frames (LAC ORFs) conserved in homologous loci

A clone isolated from a Drosophila auraria heat-shock cDNA library presents two long, antiparallel, coupled (LAC) open reading frames (ORFs). One strand ORF is 1,929 nucleotides long and exhibits great identity (87.5% at the nucleotide level and 94% at the amino acid level) with the hsp70 gene copies of D. melanogaster, while the second strand ORF, in antiparallel in-frame register arrangement, is 1,839 nucleotides long and exhibits 32% identity with a putative, recently identified, NAD⁺-dependent glutamate dehydrogenase (NAD⁺-GDH). The overlap of the two ORFs is 1,824 nucleotides long. Computational analysis shows that this LAC ORF arrangement is conserved in other hsp70 loci in a wide range of organisms, raising questions about possible evolutionary benefits of such a peculiar genomic organization.Correspondence to: Z.G. Scouras     

The cytidylyltransferase superfamily: Identification of the nucleotide-binding site and fold prediction

The crystal structure of glycerol-3-phosphate cytidylyltransferase from B. subtilis (TagD) is about to be solved. Here, we report a testable structure prediction based on the identification by sequence analysis of a superfamily of functionally diverse but structurally similar nucleotide-binding enzymes. We predict that TagD is a member of this family. The most conserved region in this superfamily resembles the ATP-binding HiGH motif of class I aminoacyI-tRNA synthetases. The predicted secondary structure of cytidylyltransferase and its homologues is compatible with the α/β topography of the class I aminoacyl-tRNA synthetases. The hypothesis of similarity of fold is strengthened by sequence-structure alignment and 3D model building using the known structure of tyrosyl tRNA synthetase as template. The proposed 3D model of TagD is plausible both structurally, with a well packed hydrophobic core, and functionally, as the most conserved residues cluster around the putative nucleotide binding site. If correct, the model would imply a very ancient evolutionary link between class I tRNA synthetases and the novel cytidylyltransferase superfamily. © 1995 Wiley-Liss, Inc.     

Progress of 1D protein structure prediction at last

Accuracy of predicting protein secondary structure and solvent accessibility from sequence information has been improved significantly by using information contained in multiple sequence alignments as input to a neural 'network system. For the Asilomar meeting, predictions for 13 proteins were generated automatically using the publicly available prediction method PHD. The results confirm the estimate of 72% three-state prediction accuracy. The fairly accurate predictions of secondary structure segments made the tool useful as a starting point for modeling of higher dimensional aspects of protein structure. © 1995 Wiley-Liss, Inc.     

Evidence for a functional role of RNA in centrioles.

Basal bodies, purified from Chlamydomonas and Tetrahymena, were exposed to various enzymatic treatments and then assayed for their ability to nucleate aster formation upon injection into eggs of Xenopus laevis. Untreated basal bodies injected into frog eggs act as centrioles and induce the formation of asters. The aster-inducing activity of basal bodies was eliminated by treatment with proteolytic enzymes and ribonucleases. Aster-inducing activity was not affected by DNAse and a number of other enzymes. The effect of proteolytic digestion on aster-inducing activity appeared to be directly correlated with the degree of structural damage to the basal body. Low concentrations of pancreatic ribonuclease A, ribonuclease T1, and S1 nuclease also completely abolished aster-inducing activity, although these enzymes had no effect on basal body structure. Ribonuclease-treated basal bodies remained capable of supporting microtubule elongation in vitro. Preliminary evidence indicates that basal bodies from Chlamydomonas and Tetrahymena contain about 5 x 10(-16) g of RNA which co-band with basal bodies and aster-inducing activity by equilibrium density gradient sedimentation. We conclude first, that centrioles contain RNA which is required for initiation of aster formation, and second, that the centriole activity or ability to assemble a mitotic aster is separable from the basal body activity, or ability to serve directly as a template for microtubule growth.     

Mechanism of the cardiovascular response to systemic intravenous administration of leucine-enkephalin in the conscious dog

Leucine-enkephalin (Leu⁵-ENK) (35 μg/kg) increased heart rate and mean systemic arterial blood pressure following intravenous injection into chronically-instrumented, conscious dogs. Repeated injections at five-minute intervals were not associated with a diminished response. Naloxone (1 mg/kg) pre-treatment inhibited both heart rate and blood pressure increases. Prazosin (1 mg/kg) attenuated the increase in blood pressure but did not influence the heart rate response. Propranolol (1 mg/kg) attenuated the heart rate response but not the pressor response. Clonidine (30 μg/kg) attenuated the positive chronotropic effect of Leu⁵-ENK. Atropine (1 mg/kg) plus propranolol (1 mg/kg) blocked the heart rate response but the pressor effect was still present. The attenuation of the heart rate response by propranolol and the pressor response by prazosin suggests an adrenergic component to the enkephalin response; the reduction in the heart rate response by clonidine and atropine-propranolol indicates a role for cholinergic mechanisms in the chronotropic response. Hexamethonium (10 mg/kg) blocked the heart rate response and markedly inhibited the pressor response. Vagal interruption attenuated both heart rate and blood pressure responses. It is concluded that intravenous Leu⁵-ENK stimulates afferent pathways located in fibers which are contained in the vagosympathetic trunk to reflexly increase heart rate and blood pressure.