Quantification and discovery of sequence determinants of protein-per-mRNA amount in 29 human tissues

Dynamics of Single-Cell Protein Covariation during Epithelial-Mesenchymal Transition

Physiological processes, such as the epithelial-mesenchymal transition (EMT), are mediated by changes in protein interactions. These changes may be better reflected in protein covariation within a cellular cluster than in the temporal dynamics of cluster-average protein abundance. To explore this possibility, we quantified proteins in single human cells undergoing EMT. Covariation analysis of the data revealed that functionally coherent protein clusters dynamically changed their protein-protein correlations without concomitant changes in the cluster-average protein abundance. These dynamics of protein-protein correlations were monotonic in time and delineated protein modules functioning in actin cytoskeleton organization, energy metabolism, and protein transport. These protein modules are defined by protein covariation within the same time point and cluster and, thus, reflect biological regulation masked by the cluster-average protein dynamics. Thus, protein correlation dynamics across single cells offers a window into protein regulation during physiological transitions.

Multi-Omic Analysis of Esophageal Adenocarcinoma Uncovers Candidate Therapeutic Targets and Cancer-Selective Posttranscriptional Regulation

Efforts to address the poor prognosis associated with esophageal adenocarcinoma (EAC) have been hampered by a lack of biomarkers to identify early disease and therapeutic targets. Despite extensive efforts to understand the somatic mutations associated with EAC over the past decade, a gap remains in understanding how the atlas of genomic aberrations in this cancer impacts the proteome and which somatic variants are of importance for the disease phenotype. We performed a quantitative proteomic analysis of 23 EACs and matched adjacent normal esophageal and gastric tissues. We explored the correlation of transcript and protein abundance using tissue-matched RNA-seq and proteomic data from seven patients and further integrated these data with a cohort of EAC RNA-seq data (n = 264 patients), EAC whole-genome sequencing (n = 454 patients), and external published datasets. We quantified protein expression from 5879 genes in EAC and patient-matched normal tissues. Several biomarker candidates with EAC-selective expression were identified, including the transmembrane protein GPA33. We further verified the EAC-enriched expression of GPA33 in an external cohort of 115 patients and confirm this as an attractive diagnostic and therapeutic target. To further extend the insights gained from our proteomic data, an integrated analysis of protein and RNA expression in EAC and normal tissues revealed several genes with poorly correlated protein and RNA abundance, suggesting posttranscriptional regulation of protein expression. These outlier genes, including SLC25A30, TAOK2, and AGMAT, only rarely demonstrated somatic mutation, suggesting post-transcriptional drivers for this EAC-specific phenotype. AGMAT was demonstrated to be overexpressed at the protein level in EAC compared to adjacent normal tissues with an EAC-selective, post-transcriptional mechanism of regulation of protein abundance proposed. Integrated analysis of proteome, transcriptome, and genome in EAC has revealed several genes with tumor-selective, posttranscriptional regulation of protein expression, which may be an exploitable vulnerability.

Dynamics of single-cell protein covariation during epithelial-mesenchymal transition

Physiological processes, such as epithelial-mesenchymal transition (EMT), are mediated by changes in protein interactions. These changes may be better reflected in protein covariation within cellular cluster than in the temporal dynamics of cluster-average protein abundance. To explore this possibility, we quantified proteins in single human cells undergoing EMT. Covariation analysis of the data revealed that functionally coherent protein clusters dynamically changed their protein-protein correlations without concomitant changes in cluster-average protein abundance. These dynamics of protein-protein correlations were monotonic in time and delineated protein modules functioning in actin cytoskeleton organization, energy metabolism and protein transport. These protein modules are defined by protein covariation within the same time point and cluster and thus reflect biological regulation masked by the cluster-average protein dynamics. Thus, protein correlation dynamics across single cells offer a window into protein regulation during physiological transitions.

Workability of mRNA Sequencing for Predicting Protein Abundance

Transcriptomics methods (RNA-Seq, PCR) today are more routine and reproducible than proteomics methods, i.e., both mass spectrometry and immunochemical analysis. For this reason, most scientific studies are limited to assessing the level of mRNA content. At the same time, protein content (and its post-translational status) largely determines the cell's state and behavior. Such a forced extrapolation of conclusions from the transcriptome to the proteome often seems unjustified. The ratios of "transcript-protein" pairs can vary by several orders of magnitude for different genes. As a rule, the correlation coefficient between transcriptome-proteome levels for different tissues does not exceed 0.3-0.5. Several characteristics determine the ratio between the content of mRNA and protein: among them, the rate of movement of the ribosome along the mRNA and the number of free ribosomes in the cell, the availability of tRNA, the secondary structure, and the localization of the transcript. The technical features of the experimental methods also significantly influence the levels of the transcript and protein of the corresponding gene on the outcome of the comparison. Given the above biological features and the performance of experimental and bioinformatic approaches, one may develop various models to predict proteomic profiles based on transcriptomic data. This review is devoted to the ability of RNA sequencing methods for protein abundance prediction.

Modeling and interpretation of single-cell proteogenomic data

Biological functions stem from coordinated interactions among proteins, nucleic acids and small molecules. Mass spectrometry technologies for reliable, high throughput single-cell proteomics will add a new modality to genomics and enable data-driven modeling of the molecular mechanisms coordinating proteins and nucleic acids at single-cell resolution. This promising potential requires estimating the reliability of measurements and computational analysis so that models can distinguish biological regulation from technical artifacts. We highlight different measurement modes that can support single-cell proteogenomic analysis and how to estimate their reliability. We then discuss approaches for developing both abstract and mechanistic models that aim to biologically interpret the measured differences across modalities, including specific applications to directed stem cell differentiation and to inferring protein interactions in cancer cells from the buffing of DNA copy-number variations. Single-cell proteogenomic data will support mechanistic models of direct molecular interactions that will provide generalizable and predictive representations of biological systems.

Accurate prediction of in vivo protein abundances by coupling constraint-based modelling and machine learning

Quantification of how different environmental cues affect protein allocation can provide important insights for understanding cell physiology. While absolute quantification of proteins can be obtained by resource-intensive mass-spectrometry-based technologies, prediction of protein abundances offers another way to obtain insights into protein allocation. Here we present CAMEL, a framework that couples constraint-based modelling with machine learning to predict protein abundance for any environmental condition. This is achieved by building machine learning models that leverage static features, derived from protein sequences, and condition-dependent features predicted from protein-constrained metabolic models. Our findings demonstrate that CAMEL results in excellent prediction of protein allocation in E. coli (average Pearson correlation of at least 0.9), and moderate performance in S. cerevisiae (average Pearson correlation of at least 0.5). Therefore, CAMEL outperformed contending approaches without using molecular read-outs from unseen conditions and provides a valuable tool for using protein allocation in biotechnological applications.

Proteome-wide copy-number estimation from transcriptomics

Protein copy numbers constrain systems-level properties of regulatory networks, but absolute proteomic data remain scarce compared to transcriptomics obtained by RNA sequencing. We addressed this persistent gap by relating mRNA to protein statistically using best-available data from quantitative proteomics-transcriptomics for 4366 genes in 369 cell lines. The approach starts with a central estimate of protein copy number and hierarchically appends mRNA-protein and mRNA-mRNA dependencies to define an optimal gene-specific model that links mRNAs to protein. For dozens of independent cell lines and primary prostate samples, these protein inferences from mRNA outmatch stringent null models, a count-based protein-abundance repository, and empirical protein-to-mRNA ratios. The optimal mRNA-to-protein relationships capture biological processes along with hundreds of known protein-protein interaction complexes, suggesting mechanistic relationships are embedded. We use the method to estimate viral-receptor abundances of CD55-CXADR from human heart transcriptomes and build 1489 systems-biology models of coxsackievirus B3 infection susceptibility. When applied to 796 RNA sequencing profiles of breast cancer from The Cancer Genome Atlas, inferred copy-number estimates collectively reclassify 26% of Luminal A and 29% of Luminal B tumors. Protein-based reassignments strongly involve a pharmacologic target for luminal breast cancer (CDK4) and an α-catenin that is often undetectable at the mRNA level (CTTNA2). Thus, by adopting a gene-centered perspective of mRNA-protein covariation across different biological contexts, we achieve accuracies comparable to the technical reproducibility limits of contemporary proteomics. The collection of gene-specific models is assembled as a web tool for users seeking mRNA-guided predictions of absolute protein abundance (http://janeslab.shinyapps.io/Pinferna).

Synthetic Adrenocorticotropic Peptides Modulate the Expression Pattern of Immune Genes in Rat Brain following the Early Post-Stroke Period

Ischemic stroke is an acute local decrease in cerebral blood flow due to a thrombus or embolus. Of particular importance is the study of the genetic systems that determine the mechanisms underlying the formation and maintenance of a therapeutic window (a time interval of up to 6 h after a stroke) when effective treatment can be provided. Here, we used a transient middle cerebral artery occlusion (tMCAO) model in rats to study two synthetic derivatives of adrenocorticotropic hormone (ACTH). The first was ACTH(4-7)PGP, which is known as Semax. It is actively used as a neuroprotective drug. The second was the ACTH(6-9)PGP peptide, which is elucidated as a prospective agent only. Using RNA-Seq analysis, we revealed hundreds of ischemia-related differentially expressed genes (DEGs), as well as 131 and 322 DEGs related to the first and second peptide at 4.5 h after tMCAO, respectively, in dorsolateral areas of the frontal cortex of rats. Furthermore, we showed that both Semax and ACTH(6-9)PGP can partially prevent changes in the immune- and neurosignaling-related gene expression profiles disturbed by the action of ischemia at 4.5 h after tMCAO. However, their different actions with regard to predominantly immune-related genes were also revealed. This study gives insight into how the transcriptome depends on the variation in the structure of the related peptides, and it is valuable from the standpoint of the development of measures for early post-stroke therapy.

Dynamic Spatial-temporal Expression Ratio of X Chromosome to Autosomes but Stable Dosage Compensation in Mammals

In the evolutionary model of dosage compensation, per-allele expression level of the X chromosome has been proposed to have twofold up-regulation to compensate its dose reduction in males (XY) compared to females (XX). However, the expression regulation of X-linked genes is still controversial, and comprehensive evaluations are still lacking. By integrating multi-omics datasets in mammals, we investigated the expression ratios including X to autosomes (X:AA ratio) and X to orthologs (X:XX ratio) at the transcriptome, translatome, and proteome levels. We revealed a dynamic spatial-temporal X:AA ratio during development in humans and mice. Meanwhile, by tracing the evolution of orthologous gene expression in chickens, platypuses, and opossums, we found a stable expression ratio of X-linked genes in humans to their autosomal orthologs in other species (X:XX ≈ 1) across tissues and developmental stages, demonstrating stable dosage compensation in mammals. We also found that different epigenetic regulations contributed to the high tissue specificity and stage specificity of X-linked gene expression, thus affecting X:AA ratios. It could be concluded that the dynamics of X:AA ratios were attributed to the different gene contents and expression preferences of the X chromosome, rather than the stable dosage compensation.

Current sequence-based models capture gene expression determinants in promoters but mostly ignore distal enhancers

BACKGROUND

The largest sequence-based models of transcription control to date are obtained by predicting genome-wide gene regulatory assays across the human genome. This setting is fundamentally correlative, as those models are exposed during training solely to the sequence variation between human genes that arose through evolution, questioning the extent to which those models capture genuine causal signals.

RESULTS

Here we confront predictions of state-of-the-art models of transcription regulation against data from two large-scale observational studies and five deep perturbation assays. The most advanced of these sequence-based models, Enformer, by and large, captures causal determinants of human promoters. However, models fail to capture the causal effects of enhancers on expression, notably in medium to long distances and particularly for highly expressed promoters. More generally, the predicted impact of distal elements on gene expression predictions is small and the ability to correctly integrate long-range information is significantly more limited than the receptive fields of the models suggest. This is likely caused by the escalating class imbalance between actual and candidate regulatory elements as distance increases.

CONCLUSIONS

Our results suggest that sequence-based models have advanced to the point that in silico study of promoter regions and promoter variants can provide meaningful insights and we provide practical guidance on how to use them. Moreover, we foresee that it will require significantly more and particularly new kinds of data to train models accurately accounting for distal elements.

Network-Guided Multiomic Mapping of Aortic Valve Calcification

Despite devastating clinical sequelae of calcific aortic valve disease that range from left ventricular remodeling to arrhythmias, heart failure, and early death, the molecular insights into disease initiation and progression are limited and pharmacotherapies remain unavailable. The pathobiology of calcific aortic valve disease is complex and comprehensive studies are challenging valvular calcification is heterogeneous and occurs preferentially on the aortic surface, along a fibrocalcific spectrum. Here, we review efforts to study (epi-)genomic, transcriptomic, proteomic, and metabolomic aspects of aortic valve calcification in combination with network medicine-/systems biology-based strategies to integrate multilayered omics datasets and prioritize druggable targets for experimental validation studies. Ultimately, such holistic approach efforts may open therapeutic avenues that go beyond invasive and costly valve replacement therapy.

Using protein-per-mRNA differences among human tissues in codon optimization

BACKGROUND

Codon usage and nucleotide composition of coding sequences have profound effects on protein expression. However, while it is recognized that different tissues have distinct tRNA profiles and codon usages in their transcriptomes, the effect of tissue-specific codon optimality on protein synthesis remains elusive.

RESULTS

We leverage existing state-of-the-art transcriptomics and proteomics datasets from the GTEx project and the Human Protein Atlas to compute the protein-to-mRNA ratios of 36 human tissues. Using this as a proxy of translational efficiency, we build a machine learning model that identifies codons enriched or depleted in specific tissues. We detect two clusters of tissues with an opposite pattern of codon preferences. We then use these identified patterns for the development of CUSTOM, a codon optimizer algorithm which suggests a synonymous codon design in order to optimize protein production in a tissue-specific manner. In human cell-line models, we provide evidence that codon optimization should take into account particularities of the translational machinery of the tissues in which the target proteins are expressed and that our approach can design genes with tissue-optimized expression profiles.

CONCLUSIONS

We provide proof-of-concept evidence that codon preferences exist in tissue-specific protein synthesis and demonstrate its application to synthetic gene design. We show that CUSTOM can be of benefit in biological and biotechnological applications, such as in the design of tissue-targeted therapies and vaccines.

Single-molecule imaging reveals translation-dependent destabilization of mRNAs

The relationship between mRNA translation and decay is incompletely understood, with conflicting reports suggesting that translation can either promote decay or stabilize mRNAs. The effect of translation on mRNA decay has mainly been studied using ensemble measurements and global transcription and translation inhibitors, which can have pleiotropic effects. We developed a single-molecule imaging approach to control the translation of a specific transcript that enabled simultaneous measurement of translation and mRNA decay. Our results demonstrate that mRNA translation reduces mRNA stability, and mathematical modeling suggests that this process is dependent on ribosome flux. Furthermore, our results indicate that miRNAs mediate efficient degradation of both translating and non-translating target mRNAs and reveal a predominant role for mRNA degradation in miRNA-mediated regulation. Simultaneous observation of translation and decay of single mRNAs provides a framework to directly study how these processes are interconnected in cells.

Benchmarking commonly used software suites and analysis workflows for DIA proteomics and phosphoproteomics

A plethora of software suites and multiple classes of spectral libraries have been developed to enhance the depth and robustness of data-independent acquisition (DIA) data processing. However, how the combination of a DIA software tool and a spectral library impacts the outcome of DIA proteomics and phosphoproteomics data analysis has been rarely investigated using benchmark data that mimics biological complexity. In this study, we create DIA benchmark data sets simulating the regulation of thousands of proteins in a complex background, which are collected on both an Orbitrap and a timsTOF instruments. We evaluate four commonly used software suites (DIA-NN, Spectronaut, MaxDIA and Skyline) combined with seven different spectral libraries in global proteome analysis. Moreover, we assess their performances in analyzing phosphopeptide standards and TNF-α-induced phosphoproteome regulation. Our study provides a practical guidance on how to construct a robust data analysis pipeline for different proteomics studies implementing the DIA technique.

Comparative Proteomic and Transcriptomic Analysis of the Impact of Androgen Stimulation and Darolutamide Inhibition

Several inhibitors of androgen receptor (AR) function are approved for prostate cancer treatment, and their impact on gene transcription has been described. However, the ensuing effects at the protein level are far less well understood. We focused on the AR signaling inhibitor darolutamide and confirmed its strong AR binding and antagonistic activity using the high throughput cellular thermal shift assay (CETSA HT). Then, we generated comprehensive, quantitative proteomic data from the androgen-sensitive prostate cancer cell line VCaP and compared them to transcriptomic data. Following treatment with the synthetic androgen R1881 and darolutamide, global mass spectrometry-based proteomics and label-free quantification were performed. We found a generally good agreement between proteomic and transcriptomic data upon androgen stimulation and darolutamide inhibition. Similar effects were found both for the detected expressed genes and their protein products as well as for the corresponding biological programs. However, in a few instances there was a discrepancy in the magnitude of changes induced on gene expression levels compared to the corresponding protein levels, indicating post-transcriptional regulation of protein abundance. Chromatin immunoprecipitation DNA sequencing (ChIP-seq) and Hi-C chromatin immunoprecipitation (HiChIP) revealed the presence of androgen-activated AR-binding regions and long-distance AR-mediated loops at these genes.

Protein prediction models support widespread post-transcriptional regulation of protein abundance by interacting partners

Protein and mRNA levels correlate only moderately. The availability of proteogenomics data sets with protein and transcript measurements from matching samples is providing new opportunities to assess the degree to which protein levels in a system can be predicted from mRNA information. Here we examined the contributions of input features in protein abundance prediction models. Using large proteogenomics data from 8 cancer types within the Clinical Proteomic Tumor Analysis Consortium (CPTAC) data set, we trained models to predict the abundance of over 13,000 proteins using matching transcriptome data from up to 958 tumor or normal adjacent tissue samples each, and compared predictive performances across algorithms, data set sizes, and input features. Over one-third of proteins (4,648) showed relatively poor predictability (elastic net r ≤ 0.3) from their cognate transcripts. Moreover, we found widespread occurrences where the abundance of a protein is considerably less well explained by its own cognate transcript level than that of one or more trans locus transcripts. The incorporation of additional trans-locus transcript abundance data as input features increasingly improved the ability to predict sample protein abundance. Transcripts that contribute to non-cognate protein abundance primarily involve those encoding known or predicted interaction partners of the protein of interest, including not only large multi-protein complexes as previously shown, but also small stable complexes in the proteome with only one or few stable interacting partners. Network analysis further shows a complex proteome-wide interdependency of protein abundance on the transcript levels of multiple interacting partners. The predictive model analysis here therefore supports that protein-protein interaction including in small protein complexes exert post-transcriptional influence on proteome compositions more broadly than previously recognized. Moreover, the results suggest mRNA and protein co-expression analysis may have utility for finding gene interactions and predicting expression changes in biological systems.

Novel biochemical, structural, and systems insights into inflammatory signaling revealed by contextual interaction proteomics

Protein-protein interactions (PPIs) represent the main mode of the proteome organization in the cell. In the last decade, several large-scale representations of PPI networks have captured generic aspects of the functional organization of network components but mostly lack the context of cellular states. However, the generation of context-dependent PPI networks is essential for structural and systems-level modeling of biological processes-a goal that remains an unsolved challenge. Here we describe an experimental/computational strategy to achieve a modeling of PPIs that considers contextual information. This strategy defines the composition, stoichiometry, temporal organization, and cellular requirements for the formation of target assemblies. We used this approach to generate an integrated model of the formation principles and architecture of a large signalosome, the TNF-receptor signaling complex (TNF-RSC). Overall, we show that the integration of systems- and structure-level information provides a generic, largely unexplored link between the modular proteome and cellular function.

RNA-sequencing and mass-spectrometry proteomic time-series analysis of T-cell differentiation identified multiple splice variants models that predicted validated protein biomarkers in inflammatory diseases

Profiling of mRNA expression is an important method to identify biomarkers but complicated by limited correlations between mRNA expression and protein abundance. We hypothesised that these correlations could be improved by mathematical models based on measuring splice variants and time delay in protein translation. We characterised time-series of primary human naïve CD4+ T cells during early T helper type 1 differentiation with RNA-sequencing and mass-spectrometry proteomics. We performed computational time-series analysis in this system and in two other key human and murine immune cell types. Linear mathematical mixed time delayed splice variant models were used to predict protein abundances, and the models were validated using out-of-sample predictions. Lastly, we re-analysed RNA-seq datasets to evaluate biomarker discovery in five T-cell associated diseases, further validating the findings for multiple sclerosis (MS) and asthma. The new models significantly out-performing models not including the usage of multiple splice variants and time delays, as shown in cross-validation tests. Our mathematical models provided more differentially expressed proteins between patients and controls in all five diseases. Moreover, analysis of these proteins in asthma and MS supported their relevance. One marker, sCD27, was validated in MS using two independent cohorts for evaluating response to treatment and disease prognosis. In summary, our splice variant and time delay models substantially improved the prediction of protein abundance from mRNA expression in three different immune cell types. The models provided valuable biomarker candidates, which were further validated in MS and asthma.

Comparative Use of Contralateral and Sham-Operated Controls Reveals Traces of a Bilateral Genetic Response in the Rat Brain after Focal Stroke

Ischemic stroke is a multifactorial disease with a complex etiology and global consequences. Model animals are widely used in stroke studies. Various controls, either brain samples from sham-operated (SO) animals or symmetrically located brain samples from the opposite (contralateral) hemisphere (CH), are often used to analyze the processes in the damaged (ipsilateral) hemisphere (IH) after focal stroke. However, previously, it was shown that focal ischemia can lead to metabolic and transcriptomic changes not only in the IH but also in the CH. Here, using a transient middle cerebral artery occlusion (tMCAO) model and genome-wide RNA sequencing, we identified 1941 overlapping differentially expressed genes (DEGs) with a cutoff value >1.5 and Padj < 0.05 that reflected the general transcriptome response of IH subcortical cells at 24 h after tMCAO using both SO and CH controls. Concomitantly, 861 genes were differentially expressed in IH vs. SO, whereas they were not vs. the CH control. Furthermore, they were associated with apoptosis, the cell cycle, and neurotransmitter responses. In turn, we identified 221 DEGs in IH vs. CH, which were non-DEGs vs. the SO control. Moreover, they were predominantly associated with immune-related response. We believe that both sets of non-overlapping genes recorded transcriptome changes in IH cells associated with transhemispheric differences after focal cerebral ischemia. Thus, the specific response of the CH transcriptome should be considered when using it as a control in studies of target brain regions in diseases that induce a global bilateral genetic response, such as stroke.

Mass spectrometry-based draft of the mouse proteome

The laboratory mouse ranks among the most important experimental systems for biomedical research and molecular reference maps of such models are essential informational tools. Here, we present a quantitative draft of the mouse proteome and phosphoproteome constructed from 41 healthy tissues and several lines of analyses exemplify which insights can be gleaned from the data. For instance, tissue- and cell-type resolved profiles provide protein evidence for the expression of 17,000 genes, thousands of isoforms and 50,000 phosphorylation sites in vivo. Proteogenomic comparison of mouse, human and Arabidopsis reveal common and distinct mechanisms of gene expression regulation and, despite many similarities, numerous differentially abundant orthologs that likely serve species-specific functions. We leverage the mouse proteome by integrating phenotypic drug (n > 400) and radiation response data with the proteomes of 66 pancreatic ductal adenocarcinoma (PDAC) cell lines to reveal molecular markers for sensitivity and resistance. This unique atlas complements other molecular resources for the mouse and can be explored online via ProteomicsDB and PACiFIC.

Tissue-specific multi-omics analysis of atrial fibrillation

Genome-wide association studies (GWAS) for atrial fibrillation (AF) have uncovered numerous disease-associated variants. Their underlying molecular mechanisms, especially consequences for mRNA and protein expression remain largely elusive. Thus, refined multi-omics approaches are needed for deciphering the underlying molecular networks. Here, we integrate genomics, transcriptomics, and proteomics of human atrial tissue in a cross-sectional study to identify widespread effects of genetic variants on both transcript (cis-eQTL) and protein (cis-pQTL) abundance. We further establish a novel targeted trans-QTL approach based on polygenic risk scores to determine candidates for AF core genes. Using this approach, we identify two trans-eQTLs and five trans-pQTLs for AF GWAS hits, and elucidate the role of the transcription factor NKX2-5 as a link between the GWAS SNP rs9481842 and AF. Altogether, we present an integrative multi-omics method to uncover trans-acting networks in small datasets and provide a rich resource of atrial tissue-specific regulatory variants for transcript and protein levels for cardiovascular disease gene prioritization.

Uncovering the impacts of alternative splicing on the proteome with current omics techniques

The high‐throughput sequencing of cellular RNAs has underscored a broad effect of isoform diversification through alternative splicing on the transcriptome. Moreover, the differential production of transcript isoforms from gene loci has been recognized as a critical mechanism in cell differentiation, organismal development, and disease. Yet, the extent of the impact of alternative splicing on protein production and cellular function remains a matter of debate. Multiple experimental and computational approaches have been developed in recent years to address this question. These studies have unveiled how molecular changes at different steps in the RNA processing pathway can lead to differences in protein production and have functional effects. New and emerging experimental technologies open exciting new opportunities to develop new methods to fully establish the connection between messenger RNA expression and protein production and to further investigate how RNA variation impacts the proteome and cell function.

This article is categorized under:

RNA Processing > Splicing Regulation/Alternative Splicing

Translation > Regulation

RNA Evolution and Genomics > Computational Analyses of RNA

Analysis of eukaryotic lincRNA sequences indicates signatures of hindered translation linked to selection pressure

Long intergenic noncoding RNAs (lincRNAs) represent a large fraction of transcribed loci in eukaryotic genomes. Although classified as noncoding, most lincRNAs contain open reading frames (ORFs), and it remains unclear why cytoplasmic lincRNAs are not or very inefficiently translated. Here, we analyzed signatures of hindered translation in lincRNA sequences from five eukaryotes, covering a range of natural selection pressures. In fission yeast and Caenorhabditis elegans, that is, species under strong selection, we detected significantly shorter ORFs, a suboptimal sequence context around start codons for translation initiation, and trinucleotides (“codons”) corresponding to less abundant tRNAs than for neutrally evolving control sequences, likely impeding translation elongation. For human, we detected signatures for cell-type-specific hindrance of lincRNA translation, in particular codons in abundant cytoplasmic lincRNAs corresponding to lower expressed tRNAs than control codons, in three out of five human cell lines. We verified that varying tRNA expression levels between cell lines are reflected in the amount of ribosomes bound to cytoplasmic lincRNAs in each cell line. We further propose that codons at ORF starts are particularly important for reducing ribosome-binding to cytoplasmic lincRNA ORFs. Altogether, our analyses indicate that in species under stronger selection lincRNAs evolved sequence features generally hindering translation and support cell-type-specific hindrance of translation efficiency in human lincRNAs. The sequence signatures we have identified may improve predicting peptide-coding and genuine noncoding lincRNAs in a cell type.

A Standardized Brain Molecular Atlas: A Resource for Systems Modeling and Simulation

Accurate molecular concentrations are essential for reliable analyses of biochemical networks and the creation of predictive models for molecular and systems biology, yet protein and metabolite concentrations used in such models are often poorly constrained or irreproducible. Challenges of using data from different sources include conflicts in nomenclature and units, as well as discrepancies in experimental procedures, data processing and implementation of the model. To obtain a consistent estimate of protein and metabolite levels, we integrated and normalized data from a large variety of sources to calculate Adjusted Molecular Concentrations. We found a high degree of reproducibility and consistency of many molecular species across brain regions and cell types, consistent with tight homeostatic regulation. We demonstrated the value of this normalization with differential protein expression analyses related to neurodegenerative diseases, brain regions and cell types. We also used the results in proof-of-concept simulations of brain energy metabolism. The standardized Brain Molecular Atlas overcomes the obstacles of missing or inconsistent data to support systems biology research and is provided as a resource for biomolecular modeling.

Sequence determinants as key regulators in gene expression of T cells

T cell homeostasis, T cell differentiation, and T cell effector function rely on the constant fine-tuning of gene expression. To alter the T cell state, substantial remodeling of the proteome is required. This remodeling depends on the intricate interplay of regulatory mechanisms, including post-transcriptional gene regulation. In this review, we discuss how the sequence of a transcript influences these post-transcriptional events. In particular, we review how sequence determinants such as sequence conservation, GC content, and chemical modifications define the levels of the mRNA and the protein in a T cell. We describe the effect of different forms of alternative splicing on mRNA expression and protein production, and their effect on subcellular localization. In addition, we discuss the role of sequences and structures as binding hubs for miRNAs and RNA-binding proteins in T cells. The review thus highlights how the intimate interplay of post-transcriptional mechanisms dictate cellular fate decisions in T cells.

Learning the Regulatory Code of Gene Expression

Data-driven machine learning is the method of choice for predicting molecular phenotypes from nucleotide sequence, modeling gene expression events including protein-DNA binding, chromatin states as well as mRNA and protein levels. Deep neural networks automatically learn informative sequence representations and interpreting them enables us to improve our understanding of the regulatory code governing gene expression. Here, we review the latest developments that apply shallow or deep learning to quantify molecular phenotypes and decode the cis-regulatory grammar from prokaryotic and eukaryotic sequencing data. Our approach is to build from the ground up, first focusing on the initiating protein-DNA interactions, then specific coding and non-coding regions, and finally on advances that combine multiple parts of the gene and mRNA regulatory structures, achieving unprecedented performance. We thus provide a quantitative view of gene expression regulation from nucleotide sequence, concluding with an information-centric overview of the central dogma of molecular biology.

Spatially Enriched Paralog Rearrangements Argue Functionally Diverse Ribosomes Arise during Cold Acclimation in Arabidopsis

Ribosome biogenesis is essential for plants to successfully acclimate to low temperature. Without dedicated steps supervising the 60S large subunits (LSUs) maturation in the cytosol, e.g., Rei-like (REIL) factors, plants fail to accumulate dry weight and fail to grow at suboptimal low temperatures. Around REIL, the final 60S cytosolic maturation steps include proofreading and assembly of functional ribosomal centers such as the polypeptide exit tunnel and the P-Stalk, respectively. In consequence, these ribosomal substructures and their assembly, especially during low temperatures, might be changed and provoke the need for dedicated quality controls. To test this, we blocked ribosome maturation during cold acclimation using two independent reil double mutant genotypes and tested changes in their ribosomal proteomes. Additionally, we normalized our mutant datasets using as a blank the cold responsiveness of a wild-type Arabidopsis genotype. This allowed us to neglect any reil-specific effects that may happen due to the presence or absence of the factor during LSU cytosolic maturation, thus allowing us to test for cold-induced changes that happen in the early nucleolar biogenesis. As a result, we report that cold acclimation triggers a reprogramming in the structural ribosomal proteome. The reprogramming alters the abundance of specific RP families and/or paralogs in non-translational LSU and translational polysome fractions, a phenomenon known as substoichiometry. Next, we tested whether the cold-substoichiometry was spatially confined to specific regions of the complex. In terms of RP proteoforms, we report that remodeling of ribosomes after a cold stimulus is significantly constrained to the polypeptide exit tunnel (PET), i.e., REIL factor binding and functional site. In terms of RP transcripts, cold acclimation induces changes in RP families or paralogs that are significantly constrained to the P-Stalk and the ribosomal head. The three modulated substructures represent possible targets of mechanisms that may constrain translation by controlled ribosome heterogeneity. We propose that non-random ribosome heterogeneity controlled by specialized biogenesis mechanisms may contribute to a preferential or ultimately even rigorous selection of transcripts needed for rapid proteome shifts and successful acclimation.

Predicting mean ribosome load for 5'UTR of any length using deep learning

The 5’ untranslated region plays a key role in regulating mRNA translation and consequently protein abundance. Therefore, accurate modeling of 5’UTR regulatory sequences shall provide insights into translational control mechanisms and help interpret genetic variants. Recently, a model was trained on a massively parallel reporter assay to predict mean ribosome load (MRL)—a proxy for translation rate—directly from 5’UTR sequence with a high degree of accuracy. However, this model is restricted to sequence lengths investigated in the reporter assay and therefore cannot be applied to the majority of human sequences without a substantial loss of information. Here, we introduced frame pooling, a novel neural network operation that enabled the development of an MRL prediction model for 5’UTRs of any length. Our model shows state-of-the-art performance on fixed length randomized sequences, while offering better generalization performance on longer sequences and on a variety of translation-related genome-wide datasets. Variant interpretation is demonstrated on a 5’UTR variant of the gene HBB associated with beta-thalassemia. Frame pooling could find applications in other bioinformatics predictive tasks. Moreover, our model, released open source, could help pinpoint pathogenic genetic variants.

The human genome carries a complex code. It consists of genes, which provide blueprints to assemble proteins, and regulatory elements, which control when, where, and how often particular genes are transcribed and translated into protein. To read the genome correctly and specifically to find the causes of inherited diseases, we need to be able to find and interpret these regulatory elements. Here, we focus on particular regions of the genome, the so-called 5’ untranslated regions, which play an important role in determining how often a transcribed gene is translated into protein. We develop deep learning models which can quantitatively interpret regulatory elements in human 5’ untranslated regions and use this information to predict a proxy of the translation efficiency. Our model generalizes a previous model to 5’ untranslated regions of any length, just as they are encountered in natural human genes. Because this model requires only the sequence as input, it can give estimates for the impact of mutations in the sequence, even if these particular mutations are very rare or entirely novel. Such estimates could help pinpoint mutations that disrupt the normal functioning of gene regulation, which could be used to better diagnose patients suffering from rare genetic disorders.

Proteogenomic Workflow Reveals Molecular Phenotypes Related to Breast Cancer Mammographic Appearance

Proteogenomic approaches have enabled the generat̲ion of novel information levels when compared to single omics studies although burdened by extensive experimental efforts. Here, we improved a data-independent acquisition mass spectrometry proteogenomic workflow to reveal distinct molecular features related to mammographic appearances in breast cancer. Our results reveal splicing processes detectable at the protein level and highlight quantitation and pathway complementarity between RNA and protein data. Furthermore, we confirm previously detected enrichments of molecular pathways associated with estrogen receptor-dependent activity and provide novel evidence of epithelial-to-mesenchymal activity in mammography-detected spiculated tumors. Several transcript-protein pairs displayed radically different abundances depending on the overall clinical properties of the tumor. These results demonstrate that there are differentially regulated protein networks in clinically relevant tumor subgroups, which in turn alter both cancer biology and the abundance of biomarker candidates and drug targets.

Endogenous p53 expression in human and mouse is not regulated by its 3'UTR

The TP53 gene encodes the tumor suppressor p53 which is functionally inactivated in many human cancers. Numerous studies suggested that 3′UTR-mediated p53 expression regulation plays a role in tumorigenesis and could be exploited for therapeutic purposes. However, these studies did not investigate post-transcriptional regulation of the native TP53 gene. Here, we used CRISPR/Cas9 to delete the human and mouse TP53/Trp53 3′UTRs while preserving endogenous mRNA processing. This revealed that the endogenous 3′UTR is not involved in regulating p53 mRNA or protein expression neither in steady state nor after genotoxic stress. Using reporter assays, we confirmed the previously observed repressive effects of the isolated 3′UTR. However, addition of the TP53 coding region to the reporter had a dominant negative impact on expression as its repressive effect was stronger and abrogated the contribution of the 3′UTR. Our data highlight the importance of genetic models in the validation of post-transcriptional gene regulatory effects.

Multi-Omics Approaches to Define Calcific Aortic Valve Disease Pathogenesis

Calcific aortic valve disease sits at the confluence of multiple world-wide epidemics of aging, obesity, diabetes, and renal dysfunction, and its prevalence is expected to nearly triple over the next 3 decades. This is of particularly dire clinical relevance, as calcific aortic valve disease can progress rapidly to aortic stenosis, heart failure, and eventually premature death. Unlike in atherosclerosis, and despite the heavy clinical toll, to date, no pharmacotherapy has proven effective to halt calcific aortic valve disease progression, with invasive and costly aortic valve replacement representing the only treatment option currently available. This substantial gap in care is largely because of our still-limited understanding of both normal aortic valve biology and the key regulatory mechanisms that drive disease initiation and progression. Drug discovery is further hampered by the inherent intricacy of the valvular microenvironment: a unique anatomic structure, a complex mixture of dynamic biomechanical forces, and diverse and multipotent cell populations collectively contributing to this currently intractable problem. One promising and rapidly evolving tactic is the application of multiomics approaches to fully define disease pathogenesis. Herein, we summarize the application of (epi)genomics, transcriptomics, proteomics, and metabolomics to the study of valvular heart disease. We also discuss recent forays toward the omics-based characterization of valvular (patho)biology at single-cell resolution; these efforts promise to shed new light on cellular heterogeneity in healthy and diseased valvular tissues and represent the potential to efficaciously target and treat key cell subpopulations. Last, we discuss systems biology- and network medicine-based strategies to extract meaning, mechanisms, and prioritized drug targets from multiomics datasets.

Factors affecting the rapid changes of protein under short-term heat stress

BACKGROUND

Protein content determines the state of cells. The variation in protein abundance is crucial when organisms are in the early stages of heat stress, but the reasons affecting their changes are largely unknown.

RESULTS

We quantified 47,535 mRNAs and 3742 proteins in the filling grains of wheat in two different thermal environments. The impact of mRNA abundance and sequence features involved in protein translation and degradation on protein expression was evaluated by regression analysis. Transcription, codon usage and amino acid frequency were the main drivers of changes in protein expression under heat stress, and their combined contribution explains 58.2 and 66.4% of the protein variation at 30 and 40 °C (20 °C as control), respectively. Transcription contributes more to alterations in protein content at 40 °C (31%) than at 30 °C (6%). Furthermore, the usage of codon AAG may be closely related to the rapid alteration of proteins under heat stress. The contributions of AAG were 24 and 13% at 30 and 40 °C, respectively.

CONCLUSION

In this study, we analyzed the factors affecting the changes in protein expression in the early stage of heat stress and evaluated their influence.

Natural variation of the cardiac transcriptome in humans

We investigated the gene-expression variation among humans by analysing previously published mRNA-seq and ribosome footprint profiling of heart left-ventricles from healthy donors. We ranked the genes according to their coefficient of variation values and found that the top 5% most variable genes had special features compared to the rest of the genome, such as lower mRNA levels and shorter half-lives coupled to increased translation efficiency. We observed that these genes are mostly involved with immune response and have a pleiotropic effect on disease phenotypes, indicating that asymptomatic conditions contribute to the gene expression diversity of healthy individuals.

Expanding small-molecule target space to mRNA translation regulation

Multiple layers of regulation are in place on mRNA translation to ensure that cells respond in a fast manner to environmental cues in a tissue-specific and mRNA-selective manner. Here, we discuss mRNA translation regulatory mechanisms and potential drug-intervention targets. Taking on a new scientific rational of translation regulation and consequently a new target space, we have developed a unique discovery platform that is able to identify selective small molecule drugs that affect translation of specific proteins. This approach has enabled targeting of proteins that have been considered undruggable. Our discovery platform was repeatedly utilized to identify compounds in multiple therapeutic programs, including fibrosis, oncology, anti-virals and Huntington's disease. In fibrosis, the lead compound ANI-21 has demonstrated a tissue-specific effect in lowering the translation of Collagen-I and superior efficacy over best standard of care, in both cell and animal models, mediated by a novel mechanism of action. This program is expected to enter clinical studies within 12-18 months.

Using Deep Learning to Extrapolate Protein Expression Measurements

Mass spectrometry (MS)-based quantitative proteomics experiments typically assay a subset of up to 60% of the ≈20 000 human protein coding genes. Computational methods for imputing the missing values using RNA expression data usually allow only for imputations of proteins measured in at least some of the samples. In silico methods for comprehensively estimating abundances across all proteins are still missing. Here, a novel method is proposed using deep learning to extrapolate the observed protein expression values in label-free MS experiments to all proteins, leveraging gene functional annotations and RNA measurements as key predictive attributes. This method is tested on four datasets, including human cell lines and human and mouse tissues. This method predicts the protein expression values with average R 2 scores between 0.46 and 0.54, which is significantly better than predictions based on correlations using the RNA expression data alone. Moreover, it is demonstrated that the derived models can be "transferred" across experiments and species. For instance, the model derived from human tissues gave a R 2 = 0.51 when applied to mouse tissue data. It is concluded that protein abundances generated in label-free MS experiments can be computationally predicted using functional annotated attributes and can be used to highlight aberrant protein abundance values.

Mammary transcriptome reveals cell maintenance and protein turnover support milk synthesis in early-lactation cows

A more complete understanding of the molecular mechanisms that support milk synthesis is needed to develop strategies to efficiently and sustainably meet the growing global demand for dairy products. With the postulate that coding gene transcript abundance reflects relative importance in supporting milk synthesis, we analyzed the global transcriptome of early lactation cows across magnitudes of normalized RNA-Seq read counts. Total RNA was isolated from milk samples collected from early-lactation cows (n = 6) following two treatment periods of postruminal lysine infusion of 0 or 63 g/day. Twelve libraries were prepared and sequenced on an Illumina NovaSeq6000 platform using paired end reads. Normalized read counts were averaged across both treatments, because EBseq analysis found no significant effect of lysine infusion. Approximately 10% of the total reads corresponded to 12,730 protein coding transcripts with a normalized read count mean ≥5. For functional annotation analysis, the protein coding transcripts were divided into nine categories by magnitude of reads. The 13 most abundant transcripts (≥50K reads) accounted for 67% of the 23M coding reads and included casein and whey proteins, regulators of fat synthesis and secretion, a ubiquitinating protein, and a tRNA transporter. Mammalian target of rapamycin, JAK/STAT, peroxisome proliferator-activated receptor alpha, and ubiquitin proteasome pathways were enriched with normalized reads ≥100 counts. Genes with ≤100 reads regulated tissue homeostasis and immune response. Enrichment in ontologies that reflect maintenance of translation, protein turnover, and amino acid recycling indicated that proteostatic mechanisms are central to supporting mammary function and primary milk component synthesis.

mRNAs, proteins and the emerging principles of gene expression control

Gene expression involves transcription, translation and the turnover of mRNAs and proteins. The degree to which protein abundances scale with mRNA levels and the implications in cases where this dependency breaks down remain an intensely debated topic. Here we review recent mRNA-protein correlation studies in the light of the quantitative parameters of the gene expression pathway, contextual confounders and buffering mechanisms. Although protein and mRNA levels typically show reasonable correlation, we describe how transcriptomics and proteomics provide useful non-redundant readouts. Integrating both types of data can reveal exciting biology and is an essential step in refining our understanding of the principles of gene expression control.

Mass-spectrometry-based draft of the Arabidopsis proteome

Plants are essential for life and are extremely diverse organisms with unique molecular capabilities¹. Here we present a quantitative atlas of the transcriptomes, proteomes and phosphoproteomes of 30 tissues of the model plant Arabidopsis thaliana. Our analysis provides initial answers to how many genes exist as proteins (more than 18,000), where they are expressed, in which approximate quantities (a dynamic range of more than six orders of magnitude) and to what extent they are phosphorylated (over 43,000 sites). We present examples of how the data may be used, such as to discover proteins that are translated from short open-reading frames, to uncover sequence motifs that are involved in the regulation of protein production, and to identify tissue-specific protein complexes or phosphorylation-mediated signalling events. Interactive access to this resource for the plant community is provided by the ProteomicsDB and ATHENA databases, which include powerful bioinformatics tools to explore and characterize Arabidopsis proteins, their modifications and interactions.

Translational efficiency across healthy and tumor tissues is proliferation-related

Different tissues express genes with particular codon usage and anticodon tRNA repertoires. However, the codon-anticodon co-adaptation in humans is not completely understood, nor is its effect on tissue-specific protein levels. Here, we first validated the accuracy of small RNA-seq for tRNA quantification across five human cell lines. We then analyzed the tRNA abundance of more than 8,000 tumor samples from TCGA, together with their paired mRNA-seq and proteomics data, to determine the Supply-to-Demand Adaptation. We thereby elucidate that the dynamic adaptation of the tRNA pool is largely related to the proliferative state across tissues. The distribution of such tRNA pools over the whole cellular translatome affects the subsequent translational efficiency, which functionally determines a condition-specific expression program both in healthy and tumor states. Furthermore, the aberrant translational efficiency of some codons in cancer, exemplified by ProCCA and GlyGGT, is associated with poor patient survival. The regulation of these tRNA profiles is partly explained by the tRNA gene copy numbers and their promoter DNA methylation.

Regulation of Proteins in Human Skeletal Muscle: The Role of Transcription

Regular low intensity aerobic exercise (aerobic training) provides effective protection against various metabolic disorders. Here, the roles played by transient transcriptome responses to acute exercise and by changes in baseline gene expression during up-regulation of protein content in human skeletal muscle were investigated after 2 months of aerobic training. Seven untrained males were involved in a 2 month aerobic cycling training program. Mass-spectrometry and RNA sequencing were used to evaluate proteome and transcriptome responses to training and acute exercise. We found that proteins with different functions are regulated differently at the transcriptional level; for example, a training-induced increase in the content of extracellular matrix-related proteins is regulated at the transcriptional level, while an increase in the content of mitochondrial proteins is not. An increase in the skeletal muscle content of several proteins (including mitochondrial proteins) was associated with increased protein stability, which is related to a chaperone-dependent mechanism and/or reduced regulation by proteolysis. These findings increase our understanding of the molecular mechanisms underlying regulation of protein expression in human skeletal muscle subjected to repeated stress (long term aerobic training) and may provide an opportunity to control the expression of specific proteins (e.g., extracellular matrix-related proteins, mitochondrial proteins) through physiological and/or pharmacological approaches.

New Technologies to Image Tumors

The premise of this book is the importance of the tumor microenvironment (TME). Until recently, most research on and clinical attention to cancer biology, diagnosis, and prognosis were focused on the malignant (or premalignant) cellular compartment that could be readily appreciated using standard morphology-based imaging.

Joint learning improves protein abundance prediction in cancers

BACKGROUND

The classic central dogma in biology is the information flow from DNA to mRNA to protein, yet complicated regulatory mechanisms underlying protein translation often lead to weak correlations between mRNA and protein abundances. This is particularly the case in cancer samples and when evaluating the same gene across multiple samples.

RESULTS

Here, we report a method for predicting proteome from transcriptome, using a training dataset provided by NCI-CPTAC and TCGA, consisting of transcriptome and proteome data from 77 breast and 105 ovarian cancer samples. First, we establish a generic model capturing the correlation between mRNA and protein abundance of a single gene. Second, we build a gene-specific model capturing the interdependencies among multiple genes in a regulatory network. Third, we create a cross-tissue model by joint learning the information of shared regulatory networks and pathways across cancer tissues. Our method ranked first in the NCI-CPTAC DREAM Proteogenomics Challenge, and the predictive performance is close to the accuracy of experimental replicates. Key functional pathways and network modules controlling the proteomic abundance in cancers were revealed, in particular metabolism-related genes.

CONCLUSIONS

We present a method to predict proteome from transcriptome, leveraging data from different cancer tissues to build a trans-tissue model, and suggest how to integrate information from multiple cancers to provide a foundation for further research.

The Protein-Coding Human Genome: Annotating High-Hanging Fruits

The major transcript variants of human protein-coding genes are annotated to a certain degree of accuracy combining manual curation, transcript data, and proteomics evidence. However, there is considerable disagreement on the annotation of about 2000 genes-they can be protein-coding, noncoding, or pseudogenes-and on the annotation of most of the predicted alternative transcripts. Pure transcriptome mapping approaches seem to be limited in discriminating functional expression from noise. These limitations have partially been overcome by dedicated algorithms to detect alternative spliced micro-exons and wobble splice variants. Recently, knowledge about splice mechanism and protein structure are incorporated into an algorithm to predict neighboring homologous exons, often spliced in a mutually exclusive manner. Predicted exons are evaluated by transcript data, structural compatibility, and evolutionary conservation, revealing hundreds of novel coding exons and splice mechanism re-assignments. The emerging human pan-genome is necessitating distinctive annotations incorporating differences between individuals and between populations.

A deep proteome and transcriptome abundance atlas of 29 healthy human tissues

Genome-, transcriptome- and proteome-wide measurements provide insights into how biological systems are regulated. However, fundamental aspects relating to which human proteins exist, where they are expressed and in which quantities are not fully understood. Therefore, we generated a quantitative proteome and transcriptome abundance atlas of 29 paired healthy human tissues from the Human Protein Atlas project representing human genes by 18,072 transcripts and 13,640 proteins including 37 without prior protein-level evidence. The analysis revealed that hundreds of proteins, particularly in testis, could not be detected even for highly expressed mRNAs, that few proteins show tissue-specific expression, that strong differences between mRNA and protein quantities within and across tissues exist and that protein expression is often more stable across tissues than that of transcripts. Only 238 of 9,848 amino acid variants found by exome sequencing could be confidently detected at the protein level showing that proteogenomics remains challenging, needs better computational methods and requires rigorous validation. Many uses of this resource can be envisaged including the study of gene/protein expression regulation and biomarker specificity evaluation.