Resolving the molecular details of proteome variation in the different tissues and organs of the human body will greatly increase our knowledge of human biology and disease. Here, we present a map of the human tissue proteome based on an integrated omics approach that involves quantitative transcriptomics at the tissue and organ level, combined with tissue microarray-based immunohistochemistry, to achieve spatial localization of proteins down to the single-cell level. Our tissue-based analysis detected more than 90% of the putative protein-coding genes. We used this approach to explore the human secretome, the membrane proteome, the druggable proteome, the cancer proteome, and the metabolic functions in 32 different tissues and organs. All the data are integrated in an interactive Web-based database that allows exploration of individual proteins, as well as navigation of global expression patterns, in all major tissues and organs in the human body.

Nonalcoholic fatty liver disease (NAFLD) refers to excess fat accumulation in the liver. In animal experiments and human kinetic study, we found that administration of combined metabolic activators (CMAs) promotes the oxidation of fat, attenuates the resulting oxidative stress, activates mitochondria, and eventually removes excess fat from the liver. Here, we tested the safety and efficacy of CMA in NAFLD patients in a placebo-controlled 10-week study. We found that CMA significantly decreased hepatic steatosis and levels of aspartate aminotransferase, alanine aminotransferase, uric acid, and creatinine, whereas found no differences on these variables in the placebo group after adjustment for weight loss. By integrating clinical data with plasma metabolomics and inflammatory proteomics as well as oral and gut metagenomic data, we revealed the underlying molecular mechanisms associated with the reduced hepatic fat and inflammation in NAFLD patients and identified the key players involved in the host-microbiome interactions. In conclusion, we showed that CMA can be used to develop a pharmacological treatment strategy in NAFLD patients.

Dietary polyunsaturated fatty acids (PUFA) are suggested to modulate immune function, but the effects of dietary fatty acids composition on gene expression patterns in immune organs have not been fully characterized. In the current study we investigated how dietary fatty acids composition affects the total transcriptome profile, and especially, immune related genes in two immune organs, spleen (SPL) and bone marrow cells (BMC). Four tissues with metabolic function, skeletal muscle (SKM), white adipose tissue (WAT), brown adipose tissue (BAT), and liver (LIV), were investigated as a comparison. Following 8 weeks on low fat diet (LFD), high fat diet (HFD) rich in saturated fatty acids (HFD-S), or HFD rich in PUFA (HFD-P), tissue transcriptomics were analyzed by microarray and metabolic health assessed by fasting blood glucose level, HOMA-IR index, oral glucose tolerance test as well as quantification of crown-like structures in WAT. HFD-P corrected the metabolic phenotype induced by HFD-S. Interestingly, SKM and BMC were relatively inert to the diets, whereas the two adipose tissues (WAT and BAT) were mainly affected by HFD per se (both HFD-S and HFD-P). In particular, WAT gene expression was driven closer to that of the immune organs SPL and BMC by HFDs. The LIV exhibited different responses to both of the HFDs. Surprisingly, the spleen showed a major response to HFD-P (82 genes differed from LFD, mostly immune genes), while it was not affected at all by HFD-S (0 genes differed from LFD). In conclusion, the quantity and composition of dietary fatty acids affected the transcriptome in distinct manners in different organs. Remarkably, dietary PUFA, but not saturated fat, prompted a specific regulation of immune related genes in the spleen, opening the possibility that PUFA can regulate immune function by influencing gene expression in this organ.

Biological networks provide new opportunities for understanding the cellular biology in both health and disease states. We generated tissue specific integrated networks (INs) for liver, muscle and adipose tissues by integrating metabolic, regulatory and protein-protein interaction networks. We also generated human co-expression networks (CNs) for 46 normal tissues and 17 cancers to explore the functional relationships between genes as well as their relationships with biological functions, and investigate the overlap between functional and physical interactions provided by CNs and INs, respectively. These networks can be employed in the analysis of omics data, provide detailed insight into disease mechanisms by identifying the key biological components and eventually can be used in the development of efficient treatment strategies. Moreover, comparative analysis of the networks may allow for the identification of tissue-specific targets that can be used in the development of drugs with the minimum toxic effect to other human tissues. These context-specific INs and CNs are presented in an interactive website http://inetmodels.com without any limitation.

Yeasts are known to have versatile metabolic traits, while how these metabolic traits have evolved has not been elucidated systematically. We performed integrative evolution analysis to investigate how genomic evolution determines trait generation by reconstructing genome-scale metabolic models (GEMs) for 332 yeasts. These GEMs could comprehensively characterize trait diversity and predict enzyme functionality, thereby signifying that sequence-level evolution has shaped reaction networks towards new metabolic functions. Strikingly, using GEMs, we can mechanistically map different evolutionary events, e.g. horizontal gene transfer and gene duplication, onto relevant subpathways to explain metabolic plasticity. This demonstrates that gene family expansion and enzyme promiscuity are prominent mechanisms for metabolic trait gains, while GEM simulations reveal that additional factors, such as gene loss from distant pathways, contribute to trait losses. Furthermore, our analysis could pinpoint to specific genes and pathways that have been under positive selection and relevant for the formulation of complex metabolic traits, i.e. thermotolerance and the Crabtree effect. Our findings illustrate how multidimensional evolution in both metabolic network structure and individual enzymes drives phenotypic variations.

Human cancer cell lines are used as important model systems to study molecular mechanisms associated with tumor growth, hereunder how genomic and biological heterogeneity found in primary tumors affect cellular phenotypes. We reconstructed Genome scale metabolic models (GEMs) for eleven cell lines based on RNA-Seq data and validated the functionality of these models with data from metabolite profiling. We used cell line-specific GEMs to analyze the differences in the metabolism of cancer cell lines, and to explore the heterogeneous expression of the metabolic subsystems. Furthermore, we predicted 85 antimetabolites that can inhibit growth of, or even kill, any of the cell lines, while at the same time not being toxic for 83 different healthy human cell types. 60 of these antimetabolites were found to inhibit growth in all cell lines. Finally, we experimentally validated one of the predicted antimetabolites using two cell lines with different phenotypic origins, and found that it is effective in inhibiting the growth of these cell lines. Using immunohistochemistry, we also showed high or moderate expression levels of proteins targeted by the validated antimetabolite. Identified anti-growth factors for inhibition of cell growth may provide leads for the development of efficient cancer treatment strategies.

Understanding the genetic regulatory code governing gene expression is an important challenge in molecular biology. However, how individual coding and non-coding regions of the gene regulatory structure interact and contribute to mRNA expression levels remains unclear. Here we apply deep learning on over 20,000 mRNA datasets to examine the genetic regulatory code controlling mRNA abundance in 7 model organisms ranging from bacteria to Human. In all organisms, we can predict mRNA abundance directly from DNA sequence, with up to 82% of the variation of transcript levels encoded in the gene regulatory structure. By searching for DNA regulatory motifs across the gene regulatory structure, we discover that motif interactions could explain the whole dynamic range of mRNA levels. Co-evolution across coding and non-coding regions suggests that it is not single motifs or regions, but the entire gene regulatory structure and specific combination of regulatory elements that define gene expression levels.

Here, we present the genome of the industrial ethanol production strain Brettanomyces bruxellensis CBS 11270. The nuclear genome was found to be diploid, containing four chromosomes with sizes of ranging from 2.2 to 4.0 Mbp. A 75 Kbp mitochondrial genome was also identified. Comparing the homologous chromosomes, we detected that 0.32% of nucleotides were polymorphic, i.e. formed single nucleotide polymorphisms (SNPs), 40.6% of them were found in coding regions (i.e. 0.13% of all nucleotides formed SNPs and were in coding regions). In addition, 8,538 indels were found. The total number of protein coding genes was 4897, of them, 4,284 were annotated on chromosomes; and the mitochondrial genome contained 18 protein coding genes. Additionally, 595 genes, which were annotated, were on contigs not associated with chromosomes. A number of genes was duplicated, most of them as tandem repeats, including a six-gene cluster located on chromosome 3. There were also examples of interchromosomal gene duplications, including a duplication of a six-gene cluster, which was found on both chromosomes 1 and 4. Gene copy number analysis suggested loss of heterozygosity for 372 genes. This may reflect adaptation to relatively harsh but constant conditions of continuous fermentation. Analysis of gene topology showed that most of these losses occurred in clusters of more than one gene, the largest cluster comprising 33 genes. Comparative analysis against the wine isolate CBS 2499 revealed 88,534 SNPs and 8,133 indels. Moreover, when the scaffolds of the CBS 2499 genome assembly were aligned against the chromosomes of CBS 11270, many of them aligned completely, some have chunks aligned to different chromosomes, and some were in fact rearranged. Our findings indicate a highly dynamic genome within the species B. bruxellensis and a tendency towards reduction of gene number in long-term continuous cultivation.

Hepatocellular carcinoma (HCC) is a deadly form of liver cancer with high mortality worldwide. Unfortunately, the large heterogeneity of this disease makes it difficult to develop effective treatment strategies. Cellular network analyses have been employed to study heterogeneity in cancer, and to identify potential therapeutic targets. However, the existing approaches do not consider metabolic growth requirements, i.e., biological network functionality, to rank candidate targets while preventing toxicity to non-cancerous tissues. Here, we developed an algorithm to overcome these issues based on integration of gene expression data, genome-scale metabolic models, network controllability, and dispensability, as well as toxicity analysis. This method thus predicts and ranks potential anticancer non-toxic controlling metabolite and gene targets. Our algorithm encompasses both objective-driven and-independent tasks, and uses network topology to finally rank the predicted therapeutic targets. We employed this algorithm to the analysis of transcriptomic data for 50 HCC patients with both cancerous and non-cancerous samples. We identified several potential targets that would prevent cell growth, including 74 anticancer metabolites, and 3 gene targets (PRKACA, PGS1, and CRLS1). The predicted anticancer metabolites showed good agreement with existing FDA-approved cancer drugs, and the 3 genes were experimentally validated by performing experiments in HepG2 and Hep3B liver cancer cell lines. Our observations indicate that our novel approach successfully identifies therapeutic targets for effective treatment of cancer. This approach may also be applied to any cancer type that has tumor and non-tumor gene or protein expression data.

A challenge in developing machine learning regression models is that it is difficult to know whether maximal performance has been reached on the test dataset, or whether further model improvement is possible. In biology, this problem is particularly pronounced as sample labels (response variables) are typically obtained through experiments and therefore have experiment noise associated with them. Such label noise puts a fundamental limit to the metrics of performance attainable by regression models on the test dataset.

The gut microbiota has been proposed as an environmental factor that promotes the progression of metabolic diseases. Here, we investigated how the gut microbiota modulates the global metabolic differences in duodenum, jejunum, ileum, colon, liver, and two white adipose tissue depots obtained from conventionally raised (CONV-R) and germ-free (GF) mice using gene expression data and tissue-specific genome-scale metabolic models (GEMs). We created a generic mouse metabolic reaction (MMR) GEM, reconstructed 28 tissue-specific GEMs based on proteomics data, and manually curated GEMs for small intestine, colon, liver, and adipose tissues. We used these functional models to determine the global metabolic differences between CONV-R and GF mice. Based on gene expression data, we found that the gut microbiota affects the host amino acid (AA) metabolism, which leads to modifications in glutathione metabolism. To validate our predictions, we measured the level of AAs and N-acetylated AAs in the hepatic portal vein of CONV-R and GF mice. Finally, we simulated the metabolic differences between the small intestine of the CONV-R and GF mice accounting for the content of the diet and relative gene expression differences. Our analyses revealed that the gut microbiota influences host amino acid and glutathione metabolism in mice.

Production of chemicals and biofuels through microbial fermentation is an economical and sustainable alternative for traditional chemical synthesis. Here we present the construction of a Saccharomyces cerevisiae platform strain for high-level production of very-long-chain fatty acid (VLCFA)-derived chemicals. Through rewiring the native fatty acid elongation system and implementing a heterologous Mycobacteria FAS I system, we establish an increased biosynthesis of VLCFAs in S. cerevisiae. VLCFAs can be selectively modified towards the fatty alcohol docosanol (C22H46O) by expressing a specific fatty acid reductase. Expression of this enzyme is shown to impair cell growth due to consumption of VLCFA-CoAs. We therefore implement a dynamic control strategy for separating cell growth from docosanol production. We successfully establish high-level and selective docosanol production of 83.5 mg l-1 in yeast. This approach will provide a universal strategy towards the production of similar high value chemicals in a more scalable, stable and sustainable manner.

Amino acids find various applications in biotechnology in view of their importance in the food, feed, pharmaceutical, and personal care industries as nutrients, additives, and drugs, respectively. For the large-scale production of amino acids, microbial cell factories are widely used and the development of amino acid-producing strains has mainly focused on prokaryotes Corynebacterium glutamicum and Escherichia coli. However, the eukaryote Saccharomyces cerevisiae is becoming an even more appealing microbial host for production of amino acids and derivatives because of its superior molecular and physiological features, such as amenable to genetic engineering and high tolerance to harsh conditions. To transform S. cerevisiae into an industrial amino acid production platform, the highly coordinated and multiple layers regulation in its amino acid metabolism should be relieved and reconstituted to optimize the metabolic flux toward synthesis of target products. This chapter describes principles, strategies, and applications of modular pathway rewiring in yeast using the engineering of l-ornithine metabolism as a paradigm. Additionally, detailed protocols for in vitro module construction and CRISPR/Cas-mediated pathway assembly are provided.

Non-alcoholic fatty liver disease (NAFLD) is a leading cause of chronic liver disease worldwide. We performed network analysis to investigate the dysregulated biological processes in the disease progression and revealed the molecular mechanism underlying NAFLD. Based on network analysis, we identified a highly conserved disease-associated gene module across three different NAFLD cohorts and highlighted the predominant role of key transcriptional regulators associated with lipid and cholesterol metabolism. In addition, we revealed the detailed metabolic differences between heterogeneous NAFLD patients through integrative systems analysis of transcriptomic data and liver-specific genome-scale metabolic model. Furthermore, we identified transcription factors (TFs), including SREBF2, HNF4A, SREBF1, YY1, and KLF13, showing regulation of hepatic expression of genes in the NAFLD-associated modules and validated the TFs using data generated from a mouse NAFLD model. In conclusion, our integrative analysis facilitates the understanding of the regulatory mechanism of these perturbed TFs and their associated biological processes.

RAVEN is a commonly used MATLAB toolbox for genome-scale metabolic model (GEM) reconstruction, curation and constraint-based modelling and simulation. Here we present RAVEN Toolbox 2.0 with major enhancements, including: (i) de novo reconstruction of GEMs based on the MetaCyc pathway database; (ii) a redesigned KEGG-based reconstruction pipeline; (iii) convergence of reconstructions from various sources; (iv) improved performance, usability, and compatibility with the COBRA Toolbox. Capabilities of RAVEN 2.0 are here illustrated through de novo reconstruction of GEMs for the antibiotic-producing bacterium Streptomyces coelicolor. Comparison of the automated de novo reconstructions with the iMK1208 model, a previously published high-quality S. coelicolor GEM, exemplifies that RAVEN 2.0 can capture most of the manually curated model. The generated de novo reconstruction is subsequently used to curate iMK1208 resulting in Sco4, the most comprehensive GEM of S. coelicolor, with increased coverage of both primary and secondary metabolism. This increased coverage allows the use of Sco4 to predict novel genome editing targets for optimized secondary metabolites production. As such, we demonstrate that RAVEN 2.0 can be used not only for de novo GEM reconstruction, but also for curating existing models based on up-to-date databases. Both RAVEN 2.0 and Sco4 are distributed through GitHub to facilitate usage and further development by the community (https://github.com/SysBioChalmers/RAVEN and https://github.com/SysBioChalmers/Streptomyces_coelicolor-GEM).

Temperature is a fundamental environmental factor that shapes the evolution of organisms. Learning thermal determinants of protein sequences in evolution thus has profound significance for basic biology, drug discovery, and protein engineering. Here, we use a data set of over 3 million BRENDA enzymes labeled with optimal growth temperatures (OGTs) of their source organisms to train a deep neural network model (DeepET). The protein-temperature representations learned by DeepET provide a temperature-related statistical summary of protein sequences and capture structural properties that affect thermal stability. For prediction of enzyme optimal catalytic temperatures and protein melting temperatures via a transfer learning approach, our DeepET model outperforms classical regression models trained on rationally designed features and other deep-learning-based representations. DeepET thus holds promise for understanding enzyme thermal adaptation and guiding the engineering of thermostable enzymes.

The molecular basis of how temperature affects cell metabolism has been a long-standing question in biology, where the main obstacles are the lack of high-quality data and methods to associate temperature effects on the function of individual proteins as well as to combine them at a systems level. Here we develop and apply a Bayesian modeling approach to resolve the temperature effects in genome scale metabolic models (GEM). The approach minimizes uncertainties in enzymatic thermal parameters and greatly improves the predictive strength of the GEMs. The resulting temperature constrained yeast GEM uncovers enzymes that limit growth at superoptimal temperatures, and squalene epoxidase (ERG1) is predicted to be the most rate limiting. By replacing this single key enzyme with an ortholog from a thermotolerant yeast strain, we obtain a thermotolerant strain that outgrows the wild type, demonstrating the critical role of sterol metabolism in yeast thermosensitivity. Therefore, apart from identifying thermal determinants of cell metabolism and enabling the design of thermotolerant strains, our Bayesian GEM approach facilitates modelling of complex biological systems in the absence of high-quality data and therefore shows promise for becoming a standard tool for genome scale modeling.

The prevalence of non-alcoholic fatty liver disease (NAFLD) continues to increase dramatically, and there is no approved medication for its treatment. Recently, we predicted the underlying molecular mechanisms involved in the progression of NAFLD using network analysis and identified metabolic cofactors that might be beneficial as supplements to decrease human liver fat. Here, we first assessed the tolerability of the combined metabolic cofactors including l-serine, N-acetyl-l-cysteine (NAC), nicotinamide riboside (NR), and l-carnitine by performing a 7-day rat toxicology study. Second, we performed a human calibration study by supplementing combined metabolic cofactors and a control study to study the kinetics of these metabolites in the plasma of healthy subjects with and without supplementation. We measured clinical parameters and observed no immediate side effects. Next, we generated plasma metabolomics and inflammatory protein markers data to reveal the acute changes associated with the supplementation of the metabolic cofactors. We also integrated metabolomics data using personalized genome-scale metabolic modeling and observed that such supplementation significantly affects the global human lipid, amino acid, and antioxidant metabolism. Finally, we predicted blood concentrations of these compounds during daily long-term supplementation by generating an ordinary differential equation model and liver concentrations of serine by generating a pharmacokinetic model and finally adjusted the doses of individual metabolic cofactors for future human clinical trials.

Cells maintain reserves in their metabolic and translational capacities as a strategy to quickly respond to changing environments. Here we quantify these reserves by stepwise reducing nitrogen availability in yeast steady-state chemostat cultures, imposing severe restrictions on total cellular protein and transcript content. Combining multi-omics analysis with metabolic modeling, we find that seven metabolic superpathways maintain >50% metabolic capacity in reserve, with glucose metabolism maintaining >80% reserve capacity. Cells maintain >50% reserve in translational capacity for 2490 out of 3361 expressed genes (74%), with a disproportionately large reserve dedicated to translating metabolic proteins. Finally, ribosome reserves contain up to 30% sub-stoichiometric ribosomal proteins, with activation of reserve translational capacity associated with selective upregulation of 17 ribosomal proteins. Together, our dataset provides a quantitative link between yeast physiology and cellular economics, which could be leveraged in future cell engineering through targeted proteome streamlining.

There is a strong interest in using photosynthetic cyanobacteria as production hosts for biofuels and chemicals. Recent work has shown the benefit of pathway engineering, enzyme tolerance, and co-factor usage for improving yields of fermentation products.