Enhancing the Biological Relevance of Machine Learning Classifiers for Reverse Vaccinology.

Reverse vaccinology (RV) is a bioinformatics approach that can predict antigens with protective potential from the protein coding genomes of bacterial pathogens for subunit vaccine design. RV has become firmly established following the development of the BEXSERO® vaccine against Neisseria meningitidis serogroup B. RV studies have begun to incorporate machine learning (ML) techniques to distinguish bacterial protective antigens (BPAs) from non-BPAs. This research contributes significantly to the RV field by using permutation analysis to demonstrate that a signal for protective antigens can be curated from published data. Furthermore, the effects of the following on an ML approach to RV were also assessed: nested cross-validation, balancing selection of non-BPAs for subcellular localization, increasing the training data, and incorporating greater numbers of protein annotation tools for feature generation. These enhancements yielded a support vector machine (SVM) classifier that could discriminate BPAs (n = 200) from non-BPAs (n = 200) with an area under the curve (AUC) of 0.787. In addition, hierarchical clustering of BPAs revealed that intracellular BPAs clustered separately from extracellular BPAs. However, no immediate benefit was derived when training SVM classifiers on data sets exclusively containing intra- or extracellular BPAs. In conclusion, this work demonstrates that ML classifiers have great utility in RV approaches and will lead to new subunit vaccines in the future.

Multiomics links global surfactant dysregulation with airflow obstruction and emphysema in COPD.

Rationale: Pulmonary surfactant is vital for lung homeostasis as it reduces surface tension to prevent alveolar collapse and provides essential immune-regulatory and antipathogenic functions. Previous studies demonstrated dysregulation of some individual surfactant components in COPD. We investigated relationships between COPD disease measures and dysregulation of surfactant components to gain new insights into potential disease mechanisms. Methods: Bronchoalveolar lavage proteome and lipidome were characterised in ex-smoking mild/moderate COPD subjects (n=26) and healthy ex-smoking (n=20) and never-smoking (n=16) controls using mass spectrometry. Serum surfactant protein analysis was performed. Results: Total phosphatidylcholine, phosphatidylglycerol, phosphatidylinositol, surfactant protein (SP)-B, SP-A and SP-D concentrations were lower in COPD versus controls (log2 fold change (log2FC) -2.0, -2.2, -1.5, -0.5, -0.7 and -0.5 (adjusted p<0.02), respectively) and correlated with lung function. Total phosphatidylcholine, phosphatidylglycerol, phosphatidylinositol, SP-A, SP-B, SP-D, napsin A and CD44 inversely correlated with computed tomography small airways disease measures (expiratory to inspiratory mean lung density) (r= -0.56, r= -0.58, r= -0.45, r= -0.36, r= -0.44, r= -0.37, r= -0.40 and r= -0.39 (adjusted p<0.05)). Total phosphatidylcholine, phosphatidylglycerol, phosphatidylinositol, SP-A, SP-B, SP-D and NAPSA inversely correlated with emphysema (% low-attenuation areas): r= -0.55, r= -0.61, r= -0.48, r= -0.51, r= -0.41, r= -0.31 and r= -0.34, respectively (adjusted p<0.05). Neutrophil elastase, known to degrade SP-A and SP-D, was elevated in COPD versus controls (log2FC 0.40, adjusted p=0.0390), and inversely correlated with SP-A and SP-D. Serum SP-D was increased in COPD versus healthy ex-smoking volunteers, and predicted COPD status (area under the curve 0.85). Conclusions: Using a multiomics approach, we demonstrate, for the first time, global surfactant dysregulation in COPD that was associated with emphysema, giving new insights into potential mechanisms underlying the cause or consequence of disease.

Dual RNASeq Reveals NTHi-Macrophage Transcriptomic Changes During Intracellular Persistence.

Nontypeable Haemophilus influenzae (NTHi) is a pathobiont which chronically colonises the airway of individuals with chronic respiratory disease and is associated with poor clinical outcomes. It is unclear how NTHi persists in the airway, however accumulating evidence suggests that NTHi can invade and persist within macrophages. To better understand the mechanisms of NTHi persistence within macrophages, we developed an in vitro model of NTHi intracellular persistence using human monocyte-derived macrophages (MDM). Dual RNA Sequencing was used to assess MDM and NTHi transcriptomic regulation occurring simultaneously during NTHi persistence. Analysis of the macrophage response to NTHi identified temporally regulated transcriptomic profiles, with a specific 'core' profile displaying conserved expression of genes across time points. Gene list enrichment analysis identified enrichment of immune responses in the core gene set, with KEGG pathway analysis revealing specific enrichment of intracellular immune response pathways. NTHi persistence was facilitated by modulation of bacterial metabolic, stress response and ribosome pathways. Levels of NTHi genes bioC, mepM and dps were differentially expressed by intracellular NTHi compared to planktonic NTHi, indicating that the transcriptomic adaption was distinct between the two different NTHi lifestyles. Overall, this study provides crucial insights into the transcriptomic adaptations facilitating NTHi persistence within macrophages. Targeting these reported pathways with novel therapeutics to reduce NTHi burden in the airway could be an effective treatment strategy given the current antimicrobial resistance crisis and lack of NTHi vaccines.

Micro RNA Targets in HIV Latency: Insights into Novel Layers of Latency Control.

Despite the considerable progress that has been made in identifying cellular factors and pathways that contribute to establishment and maintenance of the latent HIV reservoir, it remains the major obstacle to eradicating this virus. Most recently, noncoding genes have been implicated in regulation of HIV expression. In this study, small RNA sequencing was used to profile expression of microRNAs (miRNAs) in a primary CD4+ T cell in vitro model of HIV latency. Previously, we have shown that protein-coding genes dysregulated in this model were enriched for the p53 signaling pathway, which was confirmed experimentally. We further found a link between p53 signaling and dysregulated long noncoding RNAs. In this study, we hypothesized that miRNAs may provide an additional level of regulation of the p53 signaling pathway during HIV latency. Twenty-six miRNAs were identified to be dysregulated in our latency model. A subset of these miRNAs was validated by real-time quantitative polymerase chain reaction. Predicted messenger RNA (mRNA) targets and cellular pathways enriched for mRNA targets were identified using several analytical methods. Our analyses showed that many protein-coding genes and pathways targeted by dysregulated miRNAs have relevance to regulation of HIV expression or establishment of HIV latency. The p53 signaling pathway was found among pathways that were targeted by dysregulated miRNAs at a greater level than expected by chance. This study provides a mechanistic insight into regulation of the p53 pathway through miRNAs that may contribute to the establishment of latency.

An evaluation of different classification algorithms for protein sequence-based reverse vaccinology prediction.

Previous work has shown that proteins that have the potential to be vaccine candidates can be predicted from features derived from their amino acid sequences. In this work, we make an empirical comparison across various machine learning classifiers on this sequence-based inference problem. Using systematic cross validation on a dataset of 200 known vaccine candidates and 200 negative examples, with a set of 525 features derived from the AA sequences and feature selection applied through a greedy backward elimination approach, we show that simple classification algorithms often perform as well as more complex support vector kernel machines. The work also includes a novel cross validation applied across bacterial species, i.e. the validation proteins all come from a specific species of bacterium not represented in the training set. We termed this type of validation Leave One Bacteria Out Validation (LOBOV).

The promise of reverse vaccinology.

Reverse vaccinology (RV) is a computational approach that aims to identify putative vaccine candidates in the protein coding genome (proteome) of pathogens. RV has primarily been applied to bacterial pathogens to identify proteins that can be formulated into subunit vaccines, which consist of one or more protein antigens. An RV approach based on a filtering method has already been used to construct a subunit vaccine against Neisseria meningitidis serogroup B that is now registered in several countries (Bexsero). Recently, machine learning methods have been used to improve the ability of RV approaches to identify vaccine candidates. Further improvements related to the incorporation of epitope-binding annotation and gene expression data are discussed. In the future, it is envisaged that RV approaches will facilitate rapid vaccine design with less reliance on conventional animal testing and clinical trials in order to curb the threat of antibiotic resistance or newly emerged outbreaks of bacterial origin.