Machine learning based analysis of single-cell data reveals evidence of subject-specific single-cell gene expression profiles in acute myeloid leukaemia patients and healthy controls.

Acute Myeloid Leukaemia (AML) is characterized by uncontrolled growth of immature myeloid cells, disrupting normal blood production. Treatment typically involves chemotherapy, targeted therapy, and stem cell transplantation but many patients develop chemoresistance, leading to poor outcomes due to the disease's high heterogeneity. In this study, we used publicly available single-cell RNA sequencing data and machine learning to classify AML patients and healthy, monocytes, dendritic and progenitor cells population. We found that gene expression profiles of AML patients and healthy controls can be classified at the individual level with high accuracy (>70 %) when using progenitor cells, suggesting the existence of subject-specific single cell transcriptomics profiles. The analysis also revealed molecular determinants of patient heterogeneity (e.g. TPSD1, CT45A1, and GABRA4) which could support new strategies for patient stratification and personalized treatment in leukaemia.

Designing interpretable deep learning applications for functional genomics: a quantitative analysis.

Deep learning applications have had a profound impact on many scientific fields, including functional genomics. Deep learning models can learn complex interactions between and within omics data; however, interpreting and explaining these models can be challenging. Interpretability is essential not only to help progress our understanding of the biological mechanisms underlying traits and diseases but also for establishing trust in these model's efficacy for healthcare applications. Recognizing this importance, recent years have seen the development of numerous diverse interpretability strategies, making it increasingly difficult to navigate the field. In this review, we present a quantitative analysis of the challenges arising when designing interpretable deep learning solutions in functional genomics. We explore design choices related to the characteristics of genomics data, the neural network architectures applied, and strategies for interpretation. By quantifying the current state of the field with a predefined set of criteria, we find the most frequent solutions, highlight exceptional examples, and identify unexplored opportunities for developing interpretable deep learning models in genomics.

Normalization of gene counts affects principal components-based exploratory analysis of RNA-sequencing data.

Normalization of gene expression count data is an essential step of in the analysis of RNA-sequencing data. Its statistical analysis has been mostly addressed in the context of differential expression analysis, that is in the univariate setting. However, relationships among genes and samples are better explored and quantified using multivariate exploratory data analysis tools like Principal Component Analysis (PCA). In this study we investigate how normalization impacts the PCA model and its interpretation, considering twelve different widely used normalization methods that were applied on simulated and experimental data. Correlation patterns in the normalized data were explored using both summary statistics and Covariance Simultaneous Component Analysis. The impact of normalization on the PCA solution was assessed by exploring the model complexity, the quality of sample clustering in the low-dimensional PCA space and gene ranking in the model fit to normalized data. PCA models upon normalization were interpreted in the context gene enrichment pathway analysis. We found that although PCA score plots are often similar independently form the normalization used, biological interpretation of the models can depend heavily on the normalization method applied.

Proteomics-based identification of biomarkers reflecting endogenous and exogenous exposure to the advanced glycation end product precursor methylglyoxal in SH-SY5Y human neuroblastoma cells.

Methylglyoxal (MGO), a highly reactive precursor of advanced glycation end products, is endogenously produced and prevalent in various food products. This study aimed to characterize protein modifications in SH-SY5Y human neuroblastoma cells induced by MGO and identify potential biomarkers for its exposure and toxicity. A shot-gun proteomic analysis was applied to characterize protein modifications in cells incubated with and without exogenous MGO. Seventy-seven proteins were identified as highly susceptible to MGO modification, among which eight, including vimentin and histone H2B type 2-F, showing concentration-dependent modifications by externally added MGO, were defined as biomarkers for exogenous MGO exposure. Remarkably, up to 10 modification sites were identified on vimentin. Myosin light polypeptide 6 emerged as a biomarker for MGO toxicity, with modifications exclusively observed under cytotoxic MGO levels. Additionally, proteins like serine/threonine-protein kinase SIK2 and calcyphosin, exhibiting comparable or even higher modification levels in control compared to exogenous MGO-treated cells, were defined as biomarkers for endogenous exposure. Bioinformatics analysis revealed that motor proteins, cytoskeleton components, and glycolysis proteins were overrepresented among those highly susceptible to MGO modification. These results identify biomarkers for both endogenous and exogenous MGO exposure and provide insights into the cellular effects of endogenously formed versus externally added MGO.

Discrimination of Lipogenic or Glucogenic Diet Effects in Early-Lactation Dairy Cows Using Plasma Metabolite Abundances and Ratios in Combination with Machine Learning.

During early lactation, dairy cows have a negative energy balance since their energy demands exceed their energy intake: in this study, we aimed to investigate the association between diet and plasma metabolomics profiles and how these relate to energy unbalance of course in the early-lactation stage. Holstein-Friesian cows were randomly assigned to a glucogenic (n = 15) or lipogenic (n = 15) diet in early lactation. Blood was collected in week 2 and week 4 after calving. Plasma metabolite profiles were detected using liquid chromatography-mass spectrometry (LC-MS), and a total of 39 metabolites were identified. Two plasma metabolomic profiles were available every week for each cow. Metabolite abundance and metabolite ratios were used for the analysis using the XGboost algorithm to discriminate between diet treatment and lactation week. Using metabolite ratios resulted in better discrimination performance compared with the metabolite abundances in assigning cows to a lipogenic diet or a glucogenic diet. The quality of the discrimination of performance of lipogenic diet and glucogenic diet effects improved from 0.606 to 0.753 and from 0.696 to 0.842 in week 2 and week 4 (as measured by area under the curve, AUC), when the metabolite abundance ratios were used instead of abundances. The top discriminating ratios for diet were the ratio of arginine to tyrosine and the ratio of aspartic acid to valine in week 2 and week 4, respectively. For cows fed the lipogenic diet, choline and the ratio of creatinine to tryptophan were top features to discriminate cows in week 2 vs. week 4. For cows fed the glucogenic diet, methionine and the ratio of 4-hydroxyproline to choline were top features to discriminate dietary effects in week 2 or week 4. This study shows the added value of using metabolite abundance ratios to discriminate between lipogenic and glucogenic diet and lactation weeks in early-lactation cows when using metabolomics data. The application of this research will help to accurately regulate the nutrition of lactating dairy cows and promote sustainable agricultural development.

Network analysis of the proteome and peptidome sheds light on human milk as a biological system.

Proteins and peptides found in human milk have bioactive potential to benefit the newborn and support healthy development. Research has been carried out on the health benefits of proteins and peptides, but many questions still need to be answered about the nature of these components, how they are formed, and how they end up in the milk. This study explored and elucidated the complexity of the human milk proteome and peptidome. Proteins and peptides were analyzed with non-targeted nanoLC-Orbitrap-MS/MS in a selection of 297 milk samples from the CHILD Cohort Study. Protein and peptide abundances were determined, and a network was inferred using Gaussian graphical modeling (GGM), allowing an investigation of direct associations. This study showed that signatures of (1) specific mechanisms of transport of different groups of proteins, (2) proteolytic degradation by proteases and aminopeptidases, and (3) coagulation and complement activation are present in human milk. These results show the value of an integrated approach in evaluating large-scale omics data sets and provide valuable information for studies that aim to associate protein or peptide profiles from biofluids such as milk with specific physiological characteristics.

Corrigendum: The human milk proteome and allergy of mother and child: exploring associations with protein abundances and protein network connectivity.

[This corrects the article DOI: 10.3389/fimmu.2022.977470.].

PortPred: Exploiting deep learning embeddings of amino acid sequences for the identification of transporter proteins and their substrates.

The physiology of every living cell is regulated at some level by transporter proteins which constitute a relevant portion of membrane-bound proteins and are involved in the movement of ions, small and macromolecules across bio-membranes. The importance of transporter proteins is unquestionable. The prediction and study of previously unknown transporters can lead to the discovery of new biological pathways, drugs and treatments. Here we present PortPred, a tool to accurately identify transporter proteins and their substrate starting from the protein amino acid sequence. PortPred successfully combines pre-trained deep learning-based protein embeddings and machine learning classification approaches and outperforms other state-of-the-art methods. In addition, we present a comparison of the most promising protein sequence embeddings (Unirep, SeqVec, ProteinBERT, ESM-1b) and their performances for this specific task.

Systemic immune activation profiles in streptococcal necrotizing soft tissue infections: A prospective multicenter study.

OBJECTIVE: Early stages with streptococcal necrotizing soft tissue infections (NSTIs) are often difficult to discern from cellulitis. Increased insight into inflammatory responses in streptococcal disease may guide correct interventions and discovery of novel diagnostic targets. METHODS: Plasma levels of 37 mediators, leucocytes and CRP from 102 patients with ß-hemolytic streptococcal NSTI derived from a prospective Scandinavian multicentre study were compared to those of 23 cases of streptococcal cellulitis. Hierarchical cluster analyses were also performed. RESULTS: Differences in mediator levels between NSTI and cellulitis cases were revealed, in particular for IL-1ß, TNFα and CXCL8 (AUC >0.90). Across streptococcal NSTI etiologies, eight biomarkers separated cases with septic shock from those without, and four mediators predicted a severe outcome. CONCLUSION: Several inflammatory mediators and wider profiles were identified as potential biomarkers of NSTI. Associations of biomarker levels to type of infection and outcomes may be utilized to improve patient care and outcomes.

Exploration of Blood Metabolite Signatures of Colorectal Cancer and Polyposis through Integrated Statistical and Network Analysis.

Colorectal cancer (CRC), one of the most prevalent and deadly cancers worldwide, generally evolves from adenomatous polyps. The understanding of the molecular mechanisms underlying this pathological evolution is crucial for diagnostic and prognostic purposes. Integrative systems biology approaches offer an optimal point of view to analyze CRC and patients with polyposis. The present study analyzed the association networks constructed from a publicly available array of 113 serum metabolites measured on a cohort of 234 subjects from three groups (66 CRC patients, 76 patients with polyposis, and 92 healthy controls), which concentrations were obtained via targeted liquid chromatography-tandem mass spectrometry. In terms of architecture, topology, and connectivity, the metabolite-metabolite association network of CRC patients appears to be completely different with respect to patients with polyposis and healthy controls. The most relevant nodes in the CRC network are those related to energy metabolism. Interestingly, phenylalanine, tyrosine, and tryptophan metabolism are found to be involved in both CRC and polyposis. Our results demonstrate that the characterization of metabolite-metabolite association networks is a promising and powerful tool to investigate molecular aspects of CRC.

OrganelX web server for sub-peroxisomal and sub-mitochondrial protein localization and peroxisomal target signal detection.

We present the OrganelX e-Science Web Server that provides a user-friendly implementation of the In-Pero and In-Mito classifiers for sub-peroxisomal and sub-mitochondrial localization of peroxisomal and mitochondrial proteins and the Is-PTS1 algorithm for detecting and validating potential peroxisomal proteins carrying a PTS1 signal sequence. The OrganelX e-Science Web Server is available at https://organelx.hpc.rug.nl/fasta/.

Association of Plasma Metabolites and Lipoproteins with Rh and ABO Blood Systems in Healthy Subjects.

This study investigated the associations between the levels of 27 plasma metabolites, 114 lipoprotein parameters, determined using nuclear magnetic resonance spectroscopy, and the ABO blood groups and the Rhesus (Rh) blood system in a cohort of n = 840 Italian healthy blood donors of both sexes. We observed good multivariate discrimination between the metabolomic and lipoproteomic profiles of subjects with positive and negative Rh. In contrast, we did not observe significant discrimination for the ABO blood group pairwise comparisons, suggesting only slight metabolic differences between these group-specific metabolic profiles. We report univariate associations (P-value < 0.05) between the subfraction HDL1 related to Apo A1, the subfraction HDL2 related to cholesterol and phospholipids, and the particle number of LDL2 related to free cholesterol, cholesterol, phospholipids, and Apo B and the ABO blood groups; we observed association of the lipid main fraction LDL4 related to free cholesterol, triglycerides, and Apo B; creatine; the particle number of LDL5; the subfraction LDL5 related to Apo B; the particle number of LDL4; and the subfraction LDL4 related to Apo B with Rh blood factors. These results suggest blood group-dependent (re)shaping of lipoprotein metabolism in healthy subjects, which may provide relevant information to explain the differential susceptibility to certain diseases observed in different blood groups.

Intraspecies characterization of bacteria via evolutionary modeling of protein domains.

The ability to detect and characterize bacteria within a biological sample is crucial for the monitoring of infections and epidemics, as well as for the study of human health and its relationship with commensal microorganisms. To this aim, a commonly used technique is the 16S rRNA gene targeted sequencing. PCR-amplified 16S sequences derived from the sample of interest are usually clustered into the so-called Operational Taxonomic Units (OTUs) based on pairwise similarities. Then, representative OTU sequences are compared with reference (human-made) databases to derive their phylogeny and taxonomic classification. Here, we propose a new reference-free approach to define the phylogenetic distance between bacteria based on protein domains, which are the evolving units of proteins. We extract the protein domain profiles of 3368 bacterial genomes and we use an ecological approach to model their Relative Species Abundance distribution. Based on the model parameters, we then derive a new measurement of phylogenetic distance. Finally, we show that such model-based distance is capable of detecting differences between bacteria in cases in which the 16S rRNA-based method fails, providing a possibly complementary approach , which is particularly promising for the analysis of bacterial populations measured by shotgun sequencing.

Decision support system and outcome prediction in a cohort of patients with necrotizing soft-tissue infections.

INTRODUCTION: Necrotizing Soft Tissue Infections (NSTI) are severe infections with high mortality affecting a heterogeneous patient population. There is a need for a clinical decision support system which predicts outcomes and provides treatment recommendations early in the disease course. METHODS: To identify relevant clinical needs, interviews with eight medical professionals (surgeons, intensivists, general practitioner, emergency department physician) were conducted. This resulted in 24 unique questions. Mortality was selected as first endpoint to develop a machine learning (Random Forest) based prediction model. For this purpose, data from the prospective, international INFECT cohort (N = 409) was used. RESULTS: Applying a feature selection procedure based on an unsupervised algorithm (Boruta) to the  > 1000 variables available in INFECT, including baseline, and both NSTI specific and NSTI non-specific clinical data yielded sixteen predictive parameters available on or prior to the first day on the intensive care unit (ICU). Using these sixteen variables 30-day mortality could be accurately predicted (AUC = 0.91, 95% CI 0.88-0.96). Except for age, all variables were related to sepsis (e.g. lactate, urine production, systole). No NSTI-specific variables were identified. Predictions significantly outperformed the SOFA score(p < 0.001, AUC = 0.77, 95% CI 0.69-0.84) and exceeded but did not significantly differ from the SAPS II score (p = 0.07, AUC = 0.88, 95% CI 0.83-0.92). The developed model proved to be stable with AUC  > 0.8 in case of high rates of missing data (50% missing) or when only using very early (<1 h) available variables. CONCLUSIONS: This study shows that mortality can be accurately predicted using a machine learning model. It lays the foundation for a more extensive, multi-endpoint clinical decision support system in which ultimately other outcomes and clinical questions (risk for septic shock, AKI, causative microbe) will be included.

The human milk proteome and allergy of mother and child: Exploring associations with protein abundances and protein network connectivity.

Background: The human milk proteome comprises a vast number of proteins with immunomodulatory functions, but it is not clear how this relates to allergy of the mother or allergy development in the breastfed infant. This study aimed to explore the relation between the human milk proteome and allergy of both mother and child. Methods: Proteins were analyzed in milk samples from a subset of 300 mother-child dyads from the Canadian CHILD Cohort Study, selected based on maternal and child allergy phenotypes. For this selection, the definition of "allergy" included food allergy, eczema, allergic rhinitis, and asthma. Proteins were analyzed with non-targeted shotgun proteomics using filter-aided sample preparation (FASP) and nanoLC-Orbitrap-MS/MS. Protein abundances, based on label-free quantification, were compared using multiple statistical approaches, including univariate, multivariate, and network analyses. Results: Using univariate analysis, we observed a trend that milk for infants who develop an allergy by 3 years of age contains higher abundances of immunoglobulin chains, irrespective of the allergy status of the mother. This observation suggests a difference in the milk's immunological potential, which might be related to the development of the infant's immune system. Furthermore, network analysis showed overall increased connectivity of proteins in the milk of allergic mothers and milk for infants who ultimately develop an allergy. This difference in connectivity was especially noted for proteins involved in the protein translation machinery and may be due to the physiological status of the mother, which is reflected in the interconnectedness of proteins in her milk. In addition, it was shown that network analysis complements the other methods for data analysis by revealing complex associations between the milk proteome and mother-child allergy status. Conclusion: Together, these findings give new insights into how the human milk proteome, through differences in the abundance of individual proteins and protein-protein associations, relates to the allergy status of mother and child. In addition, these results inspire new research directions into the complex interplay of the mother-milk-infant triad and allergy.

A large-scale analysis of codon usage bias in 4868 bacterial genomes shows association of codon adaptation index with GC content, protein functional domains and bacterial phenotypes.

Multiple synonymous codons code for the same amino acid, resulting in the degeneracy of the genetic code and in the preferred used of some codons called codon bias usage (CBU). We performed a large-scale analysis of codon usage bias analysing the distribution of the codon adaptation index (CAI) and the codon relative adaptiveness index (RA) in 4868 bacterial genomes. We found that CAI values differ significantly between protein functional domains and part of the protein outside domains and show how CAI, GC content and preferred usage of polymerase III alpha subunits are related. Additionally, we give evidence of the association between CAI and bacterial phenotypes.

Analysis of host-pathogen gene association networks reveals patient-specific response to streptococcal and polymicrobial necrotising soft tissue infections.

BACKGROUND: Necrotising soft tissue infections (NSTIs) are rapidly progressing bacterial infections usually caused by either several pathogens in unison (polymicrobial infections) or Streptococcus pyogenes (mono-microbial infection). These infections are rare and are associated with high mortality rates. However, the underlying pathogenic mechanisms in this heterogeneous group remain elusive. METHODS: In this study, we built interactomes at both the population and individual levels consisting of host-pathogen interactions inferred from dual RNA-Seq gene transcriptomic profiles of the biopsies from NSTI patients. RESULTS: NSTI type-specific responses in the host were uncovered. The S. pyogenes mono-microbial subnetwork was enriched with host genes annotated with involved in cytokine production and regulation of response to stress. The polymicrobial network consisted of several significant associations between different species (S. pyogenes, Porphyromonas asaccharolytica and Escherichia coli) and host genes. The host genes associated with S. pyogenes in this subnetwork were characterised by cellular response to cytokines. We further found several virulence factors including hyaluronan synthase, Sic1, Isp, SagF, SagG, ScfAB-operon, Fba and genes upstream and downstream of EndoS along with bacterial housekeeping genes interacting with the human stress and immune response in various subnetworks between host and pathogen. CONCLUSIONS: At the population level, we found aetiology-dependent responses showing the potential modes of entry and immune evasion strategies employed by S. pyogenes, congruent with general cellular processes such as differentiation and proliferation. After stratifying the patients based on the subject-specific networks to study the patient-specific response, we observed different patient groups with different collagens, cytoskeleton and actin monomers in association with virulence factors, immunogenic proteins and housekeeping genes which we utilised to postulate differing modes of entry and immune evasion for different bacteria in relationship to the patients' phenotype.

Lipid and metabolite correlation networks specific to clinical and biochemical covariate show differences associated with sexual dimorphism in a cohort of nonagenarians.

This study defines and estimates the metabolite-lipidic component association networks constructed from an array of 20 metabolites and 114 lipids identified and quantified via NMR spectroscopy in the serum of a cohort of 355 Italian nonagenarians and ultra-nonagenarian. Metabolite-lipid association networks were built for men and women and related to an array of 101 clinical and biochemical parameters, including the presence of diseases, bio-humoral parameters, familiarity diseases, drugs treatments, and risk factors. Different connectivity patterns were observed in lipids, branched chains amino acids, alanine, and ketone bodies, suggesting their association with the sex-related and sex-clinical condition-related intrinsic metabolic changes. Furthermore, our results demonstrate, using a holistic system biology approach, that the characterization of metabolic structures and their dynamic inter-connections is a promising tool to shed light on the dimorphic pathophysiological mechanisms of aging at the molecular level.

Age- and Sex-Dependent Changes of Free Circulating Blood Metabolite and Lipid Abundances, Correlations, and Ratios.

In this study, we investigated how the concentrations, pairwise correlations and ratios of 202 free circulating blood metabolites and lipids vary with age in a panel of n = 1 882 participants with an age range from 48 to 94 years. We report a statistically significant sex-dependent association with age of a panel of metabolites and lipids involving, in women, linoleic acid, α-linoleic acid, and carnitine, and, in men, monoacylglycerols and lysophosphatidylcholines. Evaluating the association of correlations among metabolites and/or lipids with age, we found that phosphatidylcholines correlations tend to have a positive trend associated with age in women, and monoacylglycerols and lysophosphatidylcholines correlations tend to have a negative trend associated with age in men. The association of ratio between molecular features with age reveals that decanoyl-l-carnitine/lysophosphatidylcholine ratio in women "decrease" with age, while l-carnitine/phosphatidylcholine and l-acetylcarnitine/phosphatidylcholine ratios in men "increase" with age. These results suggest an age-dependent remodeling of lipid metabolism that induces changes in cell membrane bilayer composition and cell cycle mechanisms. Furthermore, we conclude that lipidome is directly involved in this age-dependent differentiation. Our results demonstrate that, using a comprehensive approach focused on the changes of concentrations and relationships of blood metabolites and lipids, as expressed by their correlations and ratios, it is possible to obtain relevant information about metabolic dynamics associated with age.