Using a whole genome co-expression network to inform the functional characterisation of predicted genomic elements from Mycobacterium tuberculosis transcriptomic data.

A whole genome co-expression network was created using Mycobacterium tuberculosis transcriptomic data from publicly available RNA-sequencing experiments covering a wide variety of experimental conditions. The network includes expressed regions with no formal annotation, including putative short RNAs and untranslated regions of expressed transcripts, along with the protein-coding genes. These unannotated expressed transcripts were among the best-connected members of the module sub-networks, making up more than half of the 'hub' elements in modules that include protein-coding genes known to be part of regulatory systems involved in stress response and host adaptation. This data set provides a valuable resource for investigating the role of non-coding RNA, and conserved hypothetical proteins, in transcriptomic remodelling. Based on their connections to genes with known functional groupings and correlations with replicated host conditions, predicted expressed transcripts can be screened as suitable candidates for further experimental validation.

Challenges in defining the functional, non-coding, expressed genome of members of the Mycobacterium tuberculosis complex.

A definitive transcriptome atlas for the non-coding expressed elements of the members of the Mycobacterium tuberculosis complex (MTBC) does not exist. Incomplete lists of non-coding transcripts can be obtained for some of the reference genomes (e.g., M. tuberculosis H37Rv) but to what extent these transcripts have homologues in closely related species or even strains is not clear. This has implications for the analysis of transcriptomic data; non-coding parts of the transcriptome are often ignored in the absence of formal, reliable annotation. Here, we review the state of our knowledge of non-coding RNAs in pathogenic mycobacteria, emphasizing the disparities in the information included in commonly used databases. We then proceed to review ways of combining computational solutions for predicting the non-coding transcriptome with experiments that can help refine and confirm these predictions.

flexiMAP: a regression-based method for discovering differential alternative polyadenylation events in standard RNA-seq data.

MOTIVATION: We present flexible Modeling of Alternative PolyAdenylation (flexiMAP), a new beta-regression-based method implemented in R, for discovering differential alternative polyadenylation events in standard RNA-seq data. RESULTS: We show, using both simulated and real data, that flexiMAP exhibits a good balance between specificity and sensitivity and compares favourably to existing methods, especially at low fold changes. In addition, the tests on simulated data reveal some hitherto unrecognized caveats of existing methods. Importantly, flexiMAP allows modeling of multiple known covariates that often confound the results of RNA-seq data analysis. AVAILABILITY AND IMPLEMENTATION: The flexiMAP R package is available at: https://github.com/kszkop/flexiMAP. Scripts and data to reproduce the analysis in this paper are available at: https://doi.org/10.5281/zenodo.3689788. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Cell-wall synthesis and ribosome maturation are co-regulated by an RNA switch in Mycobacterium tuberculosis.

The success of Mycobacterium tuberculosis relies on the ability to switch between active growth and non-replicating persistence, associated with latent TB infection. Resuscitation promoting factors (Rpfs) are essential for the transition between these states. Rpf expression is tightly regulated as these enzymes are able to degrade the cell wall, and hence potentially lethal to the bacterium itself. We have identified a regulatory element in the 5' untranslated region (UTR) of rpfB. We demonstrate that this element is a transcriptionally regulated RNA switch/riboswitch candidate, which appears to be restricted to pathogenic mycobacteria, suggesting a role in virulence. We have used translation start site mapping to re-annotate the RpfB start codon and identified and validated a ribosome binding site that is likely to be targeted by an rpfB antisense RNA. Finally, we show that rpfB is co-transcribed with ksgA and ispE downstream. ksgA encodes a universally conserved methyltransferase involved in ribosome maturation and ispE encodes an essential kinase involved in cell wall synthesis. This arrangement implies co-regulation of resuscitation, cell wall synthesis and ribosome maturation via the RNA switch.

Untranslated Parts of Genes Interpreted: Making Heads or Tails of High-Throughput Transcriptomic Data via Computational Methods: Computational methods to discover and quantify isoforms with alternative untranslated regions.

In this review we highlight the importance of defining the untranslated parts of transcripts, and present a number of computational approaches for the discovery and quantification of alternative transcription start and poly-adenylation events in high-throughput transcriptomic data. The fate of eukaryotic transcripts is closely linked to their untranslated regions, which are determined by the position at which transcription starts and ends at a genomic locus. Although the extent of alternative transcription starts and alternative poly-adenylation sites has been revealed by sequencing methods focused on the ends of transcripts, the application of these methods is not yet widely adopted by the community. We suggest that computational methods applied to standard high-throughput technologies are a useful, albeit less accurate, alternative to the expertise-demanding 5' and 3' sequencing and they are the only option for analysing legacy transcriptomic data. We review these methods here, focusing on technical challenges and arguing for the need to include better normalization of the data and more appropriate statistical models of the expected variation in the signal.

KSHV SOX mediated host shutoff: the molecular mechanism underlying mRNA transcript processing.

Onset of the lytic phase in the KSHV life cycle is accompanied by the rapid, global degradation of host (and viral) mRNA transcripts in a process termed host shutoff. Key to this destruction is the virally encoded alkaline exonuclease SOX. While SOX has been shown to possess an intrinsic RNase activity and a potential consensus sequence for endonucleolytic cleavage identified, the structures of the RNA substrates targeted remained unclear. Based on an analysis of three reported target transcripts, we were able to identify common structures and confirm that these are indeed degraded by SOX in vitro as well as predict the presence of such elements in the KSHV pre-microRNA transcript K12-2. From these studies, we were able to determine the crystal structure of SOX productively bound to a 31 nucleotide K12-2 fragment. This complex not only reveals the structural determinants required for RNA recognition and degradation but, together with biochemical and biophysical studies, reveals distinct roles for residues implicated in host shutoff. Our results further confirm that SOX and the host exoribonuclease Xrn1 act in concert to elicit the rapid degradation of mRNA substrates observed in vivo, and that the activities of the two ribonucleases are co-ordinated.

Cmr is a redox-responsive regulator of DosR that contributes to M. tuberculosis virulence.

Mycobacterium tuberculosis (MTb) is the causative agent of pulmonary tuberculosis (TB). MTb colonizes the human lung, often entering a non-replicating state before progressing to life-threatening active infections. Transcriptional reprogramming is essential for TB pathogenesis. In vitro, Cmr (a member of the CRP/FNR super-family of transcription regulators) bound at a single DNA site to act as a dual regulator of cmr transcription and an activator of the divergent rv1676 gene. Transcriptional profiling and DNA-binding assays suggested that Cmr directly represses dosR expression. The DosR regulon is thought to be involved in establishing latent tuberculosis infections in response to hypoxia and nitric oxide. Accordingly, DNA-binding by Cmr was severely impaired by nitrosation. A cmr mutant was better able to survive a nitrosative stress challenge but was attenuated in a mouse aerosol infection model. The complemented mutant exhibited a ∼2-fold increase in cmr expression, which led to increased sensitivity to nitrosative stress. This, and the inability to restore wild-type behaviour in the infection model, suggests that precise regulation of the cmr locus, which is associated with Region of Difference 150 in hypervirulent Beijing strains of Mtb, is important for TB pathogenesis.

Prokaryotic NavMs channel as a structural and functional model for eukaryotic sodium channel antagonism.

Voltage-gated sodium channels are important targets for the development of pharmaceutical drugs, because mutations in different human sodium channel isoforms have causal relationships with a range of neurological and cardiovascular diseases. In this study, functional electrophysiological studies show that the prokaryotic sodium channel from Magnetococcus marinus (NavMs) binds and is inhibited by eukaryotic sodium channel blockers in a manner similar to the human Nav1.1 channel, despite millions of years of divergent evolution between the two types of channels. Crystal complexes of the NavMs pore with several brominated blocker compounds depict a common antagonist binding site in the cavity, adjacent to lipid-facing fenestrations proposed to be the portals for drug entry. In silico docking studies indicate the full extent of the blocker binding site, and electrophysiology studies of NavMs channels with mutations at adjacent residues validate the location. These results suggest that the NavMs channel can be a valuable tool for screening and rational design of human drugs.

Leflunomide Treatment Does Not Protect Neural Cells following Oxygen-Glucose Deprivation (OGD) In Vitro.

Neonatal hypoxia-ischemia (HI) affects 2-3 per 1000 live births in developed countries and up to 26 per 1000 live births in developing countries. It is estimated that of the 750,000 infants experiencing a hypoxic-ischemic event during birth per year, more than 400,000 will be severely affected. As treatment options are limited, rapidly identifying new therapeutic avenues is critical, and repurposing drugs already in clinical use offers a fast-track route to clinic. One emerging avenue for therapeutic intervention in neonatal HI is to target mitochondrial dysfunction, which occurs early in the development of brain injury. Mitochondrial dynamics are particularly affected, with mitochondrial fragmentation occurring at the expense of the pro-fusion protein Optic Atrophy (OPA)1. OPA1, together with mitofusins (MFN)1/2, are required for membrane fusion, and therefore, protecting their function may also safeguard mitochondrial dynamics. Leflunomide, an FDA-approved immunosuppressant, was recently identified as an activator of MFN2 with partial effects on OPA1 expression. We, therefore, treated C17.2 cells with Leflunomide before or after oxygen-glucose deprivation, an in vitro mimic of HI, to determine its efficacy as a neuroprotection and inhibitor of mitochondrial dysfunction. Leflunomide increased baseline OPA1 but not MFN2 expression in C17.2 cells. However, Leflunomide was unable to promote cell survival following OGD. Equally, there was no obvious effect on mitochondrial morphology or bioenergetics. These data align with studies suggesting that the tissue and mitochondrial protein profile of the target cell/tissue are critical for taking advantage of the therapeutic actions of Leflunomide.

Translatome analysis of Tuberous Sclerosis Complex-1 patient-derived neural progenitor cells reveal rapamycin-dependent and independent alterations.

Tuberous sclerosis complex (TSC) is an inherited neurocutaneous disorder caused by mutations in TSC1 or TSC2 genes, with patients often exhibiting neurodevelopmental (ND) manifestations termed TSC-associated neuropsychiatric disorders (TAND) including autism spectrum disorder (ASD). The hamartin-tuberin (TSC1-TSC2) protein complex inactivates mechanistic target of rapamycin complex 1 (mTORC1) signaling, leading to increased protein synthesis via inactivation of translational repressor eIF4E-binding proteins (4E-BPs). In TSC1-null neural progenitor cells (NPCs), we previously reported early ND phenotypic changes, including increased proliferation/altered neurite outgrowth, which were unaffected by mTORC1-inhibitor rapamycin. Here, using polysome-profiling to quantify translational efficiencies at a transcriptome-wide level, we observed numerous TSC1-dependent alterations in NPCs, largely recapitulated in post-mortem brains from ASD donors. Although rapamycin partially reversed TSC1-associated alterations, most neural activity/synaptic- or ASD-related genes remained insensitive but were inhibited by third-generation bi-steric, mTORC1-selective inhibitor RMC-6272, which also reversed altered ND phenotypes. Together these data reveal potential implications for treatment of TAND.

Translatome analysis of tuberous sclerosis complex 1 patient-derived neural progenitor cells reveals rapamycin-dependent and independent alterations.

BACKGROUND: Tuberous sclerosis complex (TSC) is an inherited neurocutaneous disorder caused by mutations in the TSC1 or TSC2 genes, with patients often exhibiting neurodevelopmental (ND) manifestations termed TSC-associated neuropsychiatric disorders (TAND) including autism spectrum disorder (ASD) and intellectual disability. Hamartin (TSC1) and tuberin (TSC2) proteins form a complex inhibiting mechanistic target of rapamycin complex 1 (mTORC1) signaling. Loss of TSC1 or TSC2 activates mTORC1 that, among several targets, controls protein synthesis by inhibiting translational repressor eIF4E-binding proteins. Using TSC1 patient-derived neural progenitor cells (NPCs), we recently reported early ND phenotypic changes, including increased cell proliferation and altered neurite outgrowth in TSC1-null NPCs, which were unaffected by the mTORC1 inhibitor rapamycin. METHODS: Here, we used polysome profiling, which quantifies changes in mRNA abundance and translational efficiencies at a transcriptome-wide level, to compare CRISPR-edited TSC1-null with CRISPR-corrected TSC1-WT NPCs generated from one TSC donor (one clone/genotype). To assess the relevance of identified gene expression alterations, we performed polysome profiling in postmortem brains from ASD donors and age-matched controls. We further compared effects on translation of a subset of transcripts and rescue of early ND phenotypes in NPCs following inhibition of mTORC1 using the allosteric inhibitor rapamycin versus a third-generation bi-steric, mTORC1-selective inhibitor RMC-6272. RESULTS: Polysome profiling of NPCs revealed numerous TSC1-associated alterations in mRNA translation that were largely recapitulated in human ASD brains. Moreover, although rapamycin treatment partially reversed the TSC1-associated alterations in mRNA translation, most genes related to neural activity/synaptic regulation or ASD were rapamycin-insensitive. In contrast, treatment with RMC-6272 inhibited rapamycin-insensitive translation and reversed TSC1-associated early ND phenotypes including proliferation and neurite outgrowth that were unaffected by rapamycin. CONCLUSIONS: Our work reveals ample mRNA translation alterations in TSC1 patient-derived NPCs that recapitulate mRNA translation in ASD brain samples. Further, suppression of TSC1-associated but rapamycin-insensitive translation and ND phenotypes by RMC-6272 unveils potential implications for more efficient targeting of mTORC1 as a superior treatment strategy for TAND.

A text-mining system for extracting metabolic reactions from full-text articles.

BACKGROUND: Increasingly biological text mining research is focusing on the extraction of complex relationships relevant to the construction and curation of biological networks and pathways. However, one important category of pathway - metabolic pathways - has been largely neglected.Here we present a relatively simple method for extracting metabolic reaction information from free text that scores different permutations of assigned entities (enzymes and metabolites) within a given sentence based on the presence and location of stemmed keywords. This method extends an approach that has proved effective in the context of the extraction of protein-protein interactions. RESULTS: When evaluated on a set of manually-curated metabolic pathways using standard performance criteria, our method performs surprisingly well. Precision and recall rates are comparable to those previously achieved for the well-known protein-protein interaction extraction task. CONCLUSIONS: We conclude that automated metabolic pathway construction is more tractable than has often been assumed, and that (as in the case of protein-protein interaction extraction) relatively simple text-mining approaches can prove surprisingly effective. It is hoped that these results will provide an impetus to further research and act as a useful benchmark for judging the performance of more sophisticated methods that are yet to be developed.

Visualisation of variable binding pockets on protein surfaces by probabilistic analysis of related structure sets.

BACKGROUND: Protein structures provide a valuable resource for rational drug design. For a protein with no known ligand, computational tools can predict surface pockets that are of suitable size and shape to accommodate a complementary small-molecule drug. However, pocket prediction against single static structures may miss features of pockets that arise from proteins' dynamic behaviour. In particular, ligand-binding conformations can be observed as transiently populated states of the apo protein, so it is possible to gain insight into ligand-bound forms by considering conformational variation in apo proteins. This variation can be explored by considering sets of related structures: computationally generated conformers, solution NMR ensembles, multiple crystal structures, homologues or homology models. It is non-trivial to compare pockets, either from different programs or across sets of structures. For a single structure, difficulties arise in defining particular pocket's boundaries. For a set of conformationally distinct structures the challenge is how to make reasonable comparisons between them given that a perfect structural alignment is not possible. RESULTS: We have developed a computational method, Provar, that provides a consistent representation of predicted binding pockets across sets of related protein structures. The outputs are probabilities that each atom or residue of the protein borders a predicted pocket. These probabilities can be readily visualised on a protein using existing molecular graphics software. We show how Provar simplifies comparison of the outputs of different pocket prediction algorithms, of pockets across multiple simulated conformations and between homologous structures. We demonstrate the benefits of use of multiple structures for protein-ligand and protein-protein interface analysis on a set of complexes and consider three case studies in detail: i) analysis of a kinase superfamily highlights the conserved occurrence of surface pockets at the active and regulatory sites; ii) a simulated ensemble of unliganded Bcl2 structures reveals extensions of a known ligand-binding pocket not apparent in the apo crystal structure; iii) visualisations of interleukin-2 and its homologues highlight conserved pockets at the known receptor interfaces and regions whose conformation is known to change on inhibitor binding. CONCLUSIONS: Through post-processing of the output of a variety of pocket prediction software, Provar provides a flexible approach to the analysis and visualization of the persistence or variability of pockets in sets of related protein structures.

Defining the Genes Required for Survival of Mycobacterium bovis in the Bovine Host Offers Novel Insights into the Genetic Basis of Survival of Pathogenic Mycobacteria.

Tuberculosis has severe impacts on both humans and animals. Understanding the genetic basis of survival of both Mycobacterium tuberculosis, the human-adapted species, and Mycobacterium bovis, the animal-adapted species, is crucial to deciphering the biology of both pathogens. There are several studies that identify the genes required for survival of M. tuberculosis in vivo using mouse models; however, there are currently no studies probing the genetic basis of survival of M. bovis in vivo. In this study, we utilize transposon insertion sequencing in M. bovis AF2122/97 to determine the genes required for survival in cattle. We identify genes encoding established mycobacterial virulence functions such as the ESX-1 secretion system, phthiocerol dimycocerosate (PDIM) synthesis, mycobactin synthesis, and cholesterol catabolism that are required in vivo. We show that, as in M. tuberculosis H37Rv, phoPR is required by M. bovis AF2122/97 in vivo despite the known defect in signaling through this system. Comparison to studies performed in species that are able to use carbohydrates as an energy source, such as M. bovis BCG and M. tuberculosis, suggests that there are differences in the requirement for genes involved in cholesterol import (mce4 operon) and oxidation (hsd). We report a good correlation with existing mycobacterial virulence functions but also find several novel virulence factors, including genes involved in protein mannosylation, aspartate metabolism, and glycerol-phosphate metabolism. These findings further extend our knowledge of the genetic basis of survival in vivo in bacteria that cause tuberculosis and provide insight for the development of novel diagnostics and therapeutics. IMPORTANCE This is the first report of the genetic requirements of an animal-adapted member of the Mycobacterium tuberculosis complex (MTBC) in a natural host. M. bovis has devastating impacts on cattle, and bovine tuberculosis is a considerable economic, animal welfare, and public health concern. The data highlight the importance of mycobacterial cholesterol catabolism and identify several new virulence factors. Additionally, the work informs the development of novel differential diagnostics and therapeutics for TB in both human and animal populations.

Interaction of N-methyl-2-alkenyl-4-quinolones with ATP-dependent MurE ligase of Mycobacterium tuberculosis: antibacterial activity, molecular docking and inhibition kinetics.

OBJECTIVES: The aim of this study was to comprehensively evaluate the antibacterial activity and MurE inhibition of a set of N-methyl-2-alkenyl-4-quinolones found to inhibit the growth of fast-growing mycobacteria. METHODS: Using the spot culture growth inhibition assay, MICs were determined for Mycobacterium tuberculosis H(37)Rv, Mycobacterium bovis BCG and Mycobacterium smegmatis mc(2)155. MICs were determined for Mycobacterium fortuitum, Mycobacterium phlei, methicillin-resistant Staphylococcus aureus, Escherichia coli and Pseudomonas aeruginosa using microplate dilution assays. Inhibition of M. tuberculosis MurE ligase activity was determined both by colorimetric and HPLC methods. Computational modelling and binding prediction of the quinolones in the MurE structure was performed using Glide. Kinetic experiments were conducted for understanding possible competitive relations of the quinolones with the endogenous substrates of MurE ligase. RESULTS: The novel synthetic N-methyl-2-alkenyl-4-quinolones were found to be growth inhibitors of M. tuberculosis and rapid-growing mycobacteria as well as methicillin-resistant S. aureus, while showing no inhibition for E. coli and P. aeruginosa. The quinolones were found to be inhibitory to MurE ligase of M. tuberculosis in the micromolar range (IC(50) ∼40-200 µM) when assayed either spectroscopically or by HPLC. Computational docking of the quinolones on the published M. tuberculosis MurE crystal structure suggested that the uracil recognition site is a probable binding site for the quinolones. CONCLUSIONS: N-methyl-2-alkenyl-4-quinolones are inhibitors of mycobacterial and staphylococcal growth, and show MurE ligase inhibition. Therefore, they are considered as a starting point for the development of increased affinity MurE activity disruptors.

SERAPhiC: a benchmark for in silico fragment-based drug design.

Our main objective was to compile a data set of high-quality protein-fragment complexes and make it publicly available. Once assembled, the data set was challenged using docking procedures to address the following questions: (i) Can molecular docking correctly reproduce the experimentally solved structures? (ii) How thorough must the sampling be to replicate the experimental data? (iii) Can commonly used scoring functions discriminate between the native pose and other energy minima? The data set, named SERAPhiC (Selected Fragment Protein Complexes), is publicly available in a ready-to-dock format ( http://www.iit.it/en/drug-discovery-and-development/seraphic.html ). It offers computational medicinal chemists a reliable test set for both in silico protocol assessment and software development.

Therapeutic target-site variability in α1-antitrypsin characterized at high resolution.

The intrinsic propensity of α(1)-antitrypsin to undergo conformational transitions from its metastable native state to hyperstable forms provides a motive force for its antiprotease function. However, aberrant conformational change can also occur via an intermolecular linkage that results in polymerization. This has both loss-of-function and gain-of-function effects that lead to deficiency of the protein in human circulation, emphysema and hepatic cirrhosis. One of the most promising therapeutic strategies being developed to treat this disease targets small molecules to an allosteric site in the α(1)-antitrypsin molecule. Partial filling of this site impedes polymerization without abolishing function. Drug development can be improved by optimizing data on the structure and dynamics of this site. A new 1.8 Å resolution structure of α(1)-antitrypsin demonstrates structural variability within this site, with associated fluctuations in its upper and lower entrance grooves and ligand-binding characteristics around the innermost stable enclosed hydrophobic recess. These data will allow a broader selection of chemotypes and derivatives to be tested in silico and in vitro when screening and developing compounds to modulate conformational change to block the pathological mechanism while preserving function.

Probing Differences in Gene Essentiality Between the Human and Animal Adapted Lineages of the Mycobacterium tuberculosis Complex Using TnSeq.

Members of the Mycobacterium tuberculosis complex (MTBC) show distinct host adaptations, preferences and phenotypes despite being >99% identical at the nucleic acid level. Previous studies have explored gene expression changes between the members, however few studies have probed differences in gene essentiality. To better understand the functional impacts of the nucleic acid differences between Mycobacterium bovis and Mycobacterium tuberculosis, we used the Mycomar T7 phagemid delivery system to generate whole genome transposon libraries in laboratory strains of both species and compared the essentiality status of genes during growth under identical in vitro conditions. Libraries contained insertions in 54% of possible TA sites in M. bovis and 40% of those present in M. tuberculosis, achieving similar saturation levels to those previously reported for the MTBC. The distributions of essentiality across the functional categories were similar in both species. 527 genes were found to be essential in M. bovis whereas 477 genes were essential in M. tuberculosis and 370 essential genes were common in both species. CRISPRi was successfully utilised in both species to determine the impacts of silencing genes including wag31, a gene involved in peptidoglycan synthesis and Rv2182c/Mb2204c, a gene involved in glycerophospholipid metabolism. We observed species specific differences in the response to gene silencing, with the inhibition of expression of Mb2204c in M. bovis showing significantly less growth impact than silencing its orthologue (Rv2182c) in M. tuberculosis. Given that glycerophospholipid metabolism is a validated pathway for antimicrobials, our observations suggest that target vulnerability in the animal adapted lineages cannot be assumed to be the same as the human counterpart. This is of relevance for zoonotic tuberculosis as it implies that the development of antimicrobials targeting the human adapted lineage might not necessarily be effective against the animal adapted lineage. The generation of a transposon library and the first reported utilisation of CRISPRi in M. bovis will enable the use of these tools to further probe the genetic basis of survival under disease relevant conditions.

PROCOGNATE: a cognate ligand domain mapping for enzymes.

PROCOGNATE is a database of protein cognate ligands for the domains in enzyme structures as described by CATH, SCOP and Pfam, and is available as an interactive website or a flat file. This article gives an overview of the database and its generation and presents a new website front end, as well as recent increased coverage in our dataset via inclusion of Pfam domains. We also describe navigation of the website and its features. The current version (1.3) of PROCOGNATE covers 4123, 4536, 5876 structures and 377, 326, 695 superfamilies/families in CATH, SCOP and Pfam, respectively. PROCOGNATE can be accessed at: http://www.ebi.ac.uk/thornton-srv/databases/procognate/

Mapping human metabolic pathways in the small molecule chemical space.

The work presented here is a study of human metabolic pathways, as projected in the chemical space of the small molecules they comprise, and it is composed of three parts: a) a study of the extent of clustering and overlap of these pathways in chemical space, b) the development and assessment of a statistical model for estimating the proximity to a given pathway of any small molecule, and c) the use of the above model in estimating the proximity of marketed drugs to human metabolic pathways. The distribution, overlap, and relationships of human metabolic pathways in this space are revealed using both visual and quantitative approaches. A set of selected physicochemical and topological descriptors is used to build a classifier, whose aim is to predict metabolic class and pathway membership of any small molecule. The classifier performs well for tightly clustered, isolated pathways but is, naturally, much less accurate for strongly overlapping pathways. Finally, the extent of overlap of a set of known drugs with the human metabolome is examined, and the classifier is used to predict likely cross-interactions between drugs and the major metabolic pathways in humans.