Compound Activity Prediction with Dose-Dependent Transcriptomic Profiles and Deep Learning

Predicting compound activity in assays is a long-standing challenge in drug discovery. Computational models based on compound-induced gene expression signatures from a single profiling assay have shown promise toward predicting compound activity in other, seemingly unrelated, assays. Applications of such models include predicting mechanisms-of-action (MoA) for phenotypic hits, identifying off-target activities, and identifying polypharmacologies. Here, we introduce transcriptomics-to-activity transformer (TAT) models that leverage gene expression profiles observed over compound treatment at multiple concentrations to predict the compound activity in other biochemical or cellular assays. We built TAT models based on gene expression data from a RASL-seq assay to predict the activity of 2692 compounds in 262 dose-response assays. We obtained useful models for 51% of the assays, as determined through a realistic held-out set. Prospectively, we experimentally validated the activity predictions of a TAT model in a malaria inhibition assay. With a 63% hit rate, TAT successfully identified several submicromolar malaria inhibitors. Our results thus demonstrate the potential of transcriptomic responses over compound concentration and the TAT modeling framework as a cost-efficient way to identify the bioactivities of promising compounds across many assays.

JC polyoma viruria associates with protection from chronic kidney disease independently from apolipoprotein L1 genotype in African Americans

Background

Viral infections can trigger chronic kidney disease (CKD) and the urine virome may inform risk. The Natural History of APOL1-Associated Nephropathy Study (NHAANS) reported that urine JC polyomavirus (JCPyV) associated with a lower risk of APOL1-associated nephropathy in African Americans. Herein, association was assessed between urine JCPyV with CKD in African Americans independent from the APOL1 genotype.

Methods

Quantitative polymerase chain reaction was performed for urinary detection of JCPyV and BK polyoma virus (BKPyV) in 200 newly recruited nondiabetic African Americans. A combined analysis was performed in these individuals plus 300 NHAANS participants.

Results

In the 200 new participants, urine JCPyV was present in 8.8% of CKD cases and 45.8% of nonnephropathy controls (P = 3.0 × 10-8). In those with APOL1 renal-risk genotypes, JCPyV was detected in 5.1% of cases and 40.0% of controls (P = 0.0002). In those lacking APOL1 renal-risk genotypes, JCPyV was detected in 12.2% of cases and 48.8% of controls (P = 8.5 × 10-5). BKPyV was detected in 1.3% of cases and 0.8% of controls (P  =  0.77). In a combined analysis with 300 NHAANS participants (n = 500), individuals with urine JCPyV had a 63% lower risk of CKD compared with those without urine JCPyV (odds ratio 0.37; P = 4.6 × 10-6). RNA fluorescence in situ hybridization confirmed the presence of JCPyV genomic DNA and JCPyV messenger RNA (mRNA) in nondiseased kidney.

Conclusions

Inverse relationships exist between JCPyV viruria and non-diabetic CKD. Future studies should determine whether renal inflammation associated with CKD is less permissive for JCPyV reactivation/replication or whether JCPyV is a marker of reduced host immune responsiveness that diminishes immune pathologic contributions to CKD.

Mode of Action of the Monobactam LYS228 and Mechanisms Decreasing In Vitro Susceptibility in Escherichia coli and Klebsiella pneumoniae

The monobactam scaffold is attractive for the development of new agents to treat infections caused by drug-resistant Gram-negative bacteria because it is stable to metallo-β-lactamases (MBLs). However, the clinically used monobactam aztreonam lacks stability to serine β-lactamases (SBLs) that are often coexpressed with MBLs. LYS228 is stable to MBLs and most SBLs. LYS228 bound purified Escherichia coli penicillin binding protein 3 (PBP3) similarly to aztreonam (derived acylation rate/equilibrium dissociation constant [k2/Kd ] of 367,504 s-1 M-1 and 409,229 s-1 M-1, respectively) according to stopped-flow fluorimetry. A gel-based assay showed that LYS228 bound mainly to E. coli PBP3, with weaker binding to PBP1a and PBP1b. Exposing E. coli cells to LYS228 caused filamentation consistent with impaired cell division. No single-step mutants were selected from 12 Enterobacteriaceae strains expressing different classes of β-lactamases at 8× the MIC of LYS228 (frequency, <2.5 × 10-9). At 4× the MIC, mutants were selected from 2 of 12 strains at frequencies of 1.8 × 10-7 and 4.2 × 10-9 LYS228 MICs were ≤2 μg/ml against all mutants. These frequencies compared favorably to those for meropenem and tigecycline. Mutations decreasing LYS228 susceptibility occurred in ramR and cpxA (Klebsiella pneumoniae) and baeS (E. coli and K. pneumoniae). Susceptibility of E. coli ATCC 25922 to LYS228 decreased 256-fold (MIC, 0.125 to 32 μg/ml) after 20 serial passages. Mutants accumulated mutations in ftsI (encoding the target, PBP3), baeR, acrD, envZ, sucB, and rfaI These results support the continued development of LYS228, which is currently undergoing phase II clinical trials for complicated intraabdominal infection and complicated urinary tract infection (registered at ClinicalTrials.gov under identifiers NCT03377426 and NCT03354754).

Profile hidden Markov models for the detection of viruses within metagenomic sequence data

Rapid, sensitive, and specific virus detection is an important component of clinical diagnostics. Massively parallel sequencing enables new diagnostic opportunities that complement traditional serological and PCR based techniques. While massively parallel sequencing promises the benefits of being more comprehensive and less biased than traditional approaches, it presents new analytical challenges, especially with respect to detection of pathogen sequences in metagenomic contexts. To a first approximation, the initial detection of viruses can be achieved simply through alignment of sequence reads or assembled contigs to a reference database of pathogen genomes with tools such as BLAST. However, recognition of highly divergent viral sequences is problematic, and may be further complicated by the inherently high mutation rates of some viral types, especially RNA viruses. In these cases, increased sensitivity may be achieved by leveraging position-specific information during the alignment process. Here, we constructed HMMER3-compatible profile hidden Markov models (profile HMMs) from all the virally annotated proteins in RefSeq in an automated fashion using a custom-built bioinformatic pipeline. We then tested the ability of these viral profile HMMs ("vFams") to accurately classify sequences as viral or non-viral. Cross-validation experiments with full-length gene sequences showed that the vFams were able to recall 91% of left-out viral test sequences without erroneously classifying any non-viral sequences into viral protein clusters. Thorough reanalysis of previously published metagenomic datasets with a set of the best-performing vFams showed that they were more sensitive than BLAST for detecting sequences originating from more distant relatives of known viruses. To facilitate the use of the vFams for rapid detection of remote viral homologs in metagenomic data, we provide two sets of vFams, comprising more than 4,000 vFams each, in the HMMER3 format. We also provide the software necessary to build custom profile HMMs or update the vFams as more viruses are discovered (http://derisilab.ucsf.edu/software/vFam).

Molecular insights into the biosynthesis of guadinomine: a type III secretion system inhibitor

Guadinomines are a recently discovered family of anti-infective compounds produced by Streptomyces sp. K01-0509 with a novel mode of action. With an IC(50) of 14 nM, guadinomine B is the most potent known inhibitor of the type III secretion system (TTSS) of Gram-negative bacteria. TTSS activity is required for the virulence of many pathogenic Gram-negative bacteria including Escherichia coli , Salmonella spp., Yersinia spp., Chlamydia spp., Vibrio spp., and Pseudomonas spp. The guadinomine (gdn) biosynthetic gene cluster has been cloned and sequenced and includes 26 open reading frames spanning 51.2 kb. It encodes a chimeric multimodular polyketide synthase, a nonribosomal peptide synthetase, along with enzymes responsible for the biosynthesis of the unusual aminomalonyl-acyl carrier protein extender unit and the signature carbamoylated cyclic guanidine. Its identity was established by targeted disruption of the gene cluster as well as by heterologous expression and analysis of key enzymes in the biosynthetic pathway. Identifying the guadinomine gene cluster provides critical insight into the biosynthesis of these scarce but potentially important natural products.

Merochlorins A-D, cyclic meroterpenoid antibiotics biosynthesized in divergent pathways with vanadium-dependent chloroperoxidases

Meroterpenoids are mixed polyketide-terpenoid natural products with a broad range of biological activities. Herein, we present the structures of four new meroterpenoid antibiotics, merochlorins A-D, produced by the marine bacterium Streptomyces sp. strain CNH-189, which possess novel chemical skeletons unrelated to known bacterial agents. Draft genome sequencing, mutagenesis, and heterologous biosynthesis in the genome-minimized model actinomycete Streptomyces coelicolor provided the 57.6 kb merochlorin gene cluster that contains two genes encoding rare bacterial vanadium-dependent haloperoxidase (VHPO) genes. Pathway expression of two different fosmid clones that differ largely by the presence or absence of the VHPO gene mcl40 resulted in the differential biosynthesis of merochlorin C, suggesting that Mcl40 catalyzes an unprecedented 15-membered chloronium-induced macrocyclization reaction converting merochlorin D to merochlorin C.

Virus identification in unknown tropical febrile illness cases using deep sequencing

Dengue virus is an emerging infectious agent that infects an estimated 50–100 million people annually worldwide, yet current diagnostic practices cannot detect an etiologic pathogen in ∼40% of dengue-like illnesses. Metagenomic approaches to pathogen detection, such as viral microarrays and deep sequencing, are promising tools to address emerging and non-diagnosable disease challenges. In this study, we used the Virochip microarray and deep sequencing to characterize the spectrum of viruses present in human sera from 123 Nicaraguan patients presenting with dengue-like symptoms but testing negative for dengue virus. We utilized a barcoding strategy to simultaneously deep sequence multiple serum specimens, generating on average over 1 million reads per sample. We then implemented a stepwise bioinformatic filtering pipeline to remove the majority of human and low-quality sequences to improve the speed and accuracy of subsequent unbiased database searches. By deep sequencing, we were able to detect virus sequence in 37% (45/123) of previously negative cases. These included 13 cases with Human Herpesvirus 6 sequences. Other samples contained sequences with similarity to sequences from viruses in the Herpesviridae, Flaviviridae, Circoviridae, Anelloviridae, Asfarviridae, and Parvoviridae families. In some cases, the putative viral sequences were virtually identical to known viruses, and in others they diverged, suggesting that they may derive from novel viruses. These results demonstrate the utility of unbiased metagenomic approaches in the detection of known and divergent viruses in the study of tropical febrile illness.

Dengue virus infection is a global health concern, affecting as many as 100 million people annually worldwide. A critical first step to proper treatment and control of any virus infection is a correct diagnosis. Traditional diagnostic tests for viruses depend on amplification of conserved portions of the viral genome, detection of the binding of antibodies to viral proteins, or replication of the virus in cell cultures. These methods have a major shortcoming: they are unable to detect divergent or novel viruses for which a priori sequence, serological, or cellular tropism information is not known. In our study, we use two approaches, microarrays and deep sequencing, to virus identification that are less susceptible to such shortcomings. We used these unbiased tools to search for viruses in blood collected from Nicaraguan children with clinical symptoms indicating dengue virus infection, but for whom current dengue virus detection assays yielded negative results. We were able to identify both known and divergent viruses in about one third of previously negative samples, demonstrating the utility of these approaches to detect viruses in cases of unknown dengue-like illness.

Assemblathon 1: a competitive assessment of de novo short read assembly methods

Low-cost short read sequencing technology has revolutionized genomics, though it is only just becoming practical for the high-quality de novo assembly of a novel large genome. We describe the Assemblathon 1 competition, which aimed to comprehensively assess the state of the art in de novo assembly methods when applied to current sequencing technologies. In a collaborative effort, teams were asked to assemble a simulated Illumina HiSeq data set of an unknown, simulated diploid genome. A total of 41 assemblies from 17 different groups were received. Novel haplotype aware assessments of coverage, contiguity, structure, base calling, and copy number were made. We establish that within this benchmark: (1) It is possible to assemble the genome to a high level of coverage and accuracy, and that (2) large differences exist between the assemblies, suggesting room for further improvements in current methods. The simulated benchmark, including the correct answer, the assemblies, and the code that was used to evaluate the assemblies is now public and freely available from http://www.assemblathon.org/.

Assemblathon 1: a competitive assessment of de novo short read assembly methods

Low-cost short read sequencing technology has revolutionized genomics, though it is only just becoming practical for the high-quality de novo assembly of a novel large genome. We describe the Assemblathon 1 competition, which aimed to comprehensively assess the state of the art in de novo assembly methods when applied to current sequencing technologies. In a collaborative effort, teams were asked to assemble a simulated Illumina HiSeq data set of an unknown, simulated diploid genome. A total of 41 assemblies from 17 different groups were received. Novel haplotype aware assessments of coverage, contiguity, structure, base calling, and copy number were made. We establish that within this benchmark: (1) It is possible to assemble the genome to a high level of coverage and accuracy, and that (2) large differences exist between the assemblies, suggesting room for further improvements in current methods. The simulated benchmark, including the correct answer, the assemblies, and the code that was used to evaluate the assemblies is now public and freely available from http://www.assemblathon.org/.

Validation of a diagnostic microarray for human papillomavirus: coverage of 102 genotypes

Papillomaviruses have been implicated in a variety of human diseases ranging from common warts to invasive carcinoma of the anogenital mucosa. Existing assays for genotyping human papillomavirus are restricted to a small number of types. Here, we present a comprehensive, accurate microarray strategy for detection and genotyping of 102 human papillomavirus types and validate its use in a panel of 91 anal swabs. This array has equal performance to traditional dot blot analysis with the benefits of added genotype coverage and the ability to calibrate readout over a range of sensitivity or specificity values.

Recovery of divergent avian bornaviruses from cases of proventricular dilatation disease: identification of a candidate etiologic agent

BACKGROUND

Proventricular dilatation disease (PDD) is a fatal disorder threatening domesticated and wild psittacine birds worldwide. It is characterized by lymphoplasmacytic infiltration of the ganglia of the central and peripheral nervous system, leading to central nervous system disorders as well as disordered enteric motility and associated wasting. For almost 40 years, a viral etiology for PDD has been suspected, but to date no candidate etiologic agent has been reproducibly linked to the disease.

RESULTS

Analysis of 2 PDD case-control series collected independently on different continents using a pan-viral microarray revealed a bornavirus hybridization signature in 62.5% of the PDD cases (5/8) and none of the controls (0/8). Ultra high throughput sequencing was utilized to recover the complete viral genome sequence from one of the virus-positive PDD cases. This revealed a bornavirus-like genome organization for this agent with a high degree of sequence divergence from all prior bornavirus isolates. We propose the name avian bornavirus (ABV) for this agent. Further specific ABV PCR analysis of an additional set of independently collected PDD cases and controls yielded a significant difference in ABV detection rate among PDD cases (71%, n = 7) compared to controls (0%, n = 14) (P = 0.01; Fisher's Exact Test). Partial sequence analysis of a total of 16 ABV isolates we have now recovered from these and an additional set of cases reveals at least 5 distinct ABV genetic subgroups.

CONCLUSION

These studies clearly demonstrate the existence of an avian reservoir of remarkably diverse bornaviruses and provide a compelling candidate in the search for an etiologic agent of PDD.

MicroRNA expression is required for pancreatic islet cell genesis in the mouse

OBJECTIVE

The generation of distinct cell types during the development of the pancreas depends on sequential changes in gene expression. We tested the hypothesis that microRNAs (miRNAs), which limit gene expression through posttranscriptional silencing, modulate the gene expression cascades involved in pancreas development.

RESEARCH DESIGN AND METHODS

miRNAs were cloned and sequenced from developing pancreata, and expression of a subset of these genes was tested using locked nucleic acid in situ analyses. To assess the overall contribution of miRNAs to pancreatic development, Dicer1, an enzyme required for miRNA processing, was conditionally deleted from the developing pancreas.

RESULTS

Sequencing of small RNAs identified over 125 miRNAs, including 18 novel sequences, with distinct expression domains within the developing pancreas. To test the developmental contribution of these miRNAs, we conditionally deleted the miRNA processing enzyme Dicer1 early in pancreas development. Dicer-null animals displayed gross defects in all pancreatic lineages, although the endocrine cells, and especially the insulin-producing beta-cells, were most dramatically reduced. The endocrine defect was associated with an increase in the notch-signaling target Hes1 and a reduction in the formation of endocrine cell progenitors expressing the Hes1 target gene neurogenin3.

CONCLUSIONS

The expression of a unique profile of miRNAs is required during pancreas development and is necessary for beta-cell formation.