Challenges of COVID-19 Case Forecasting in the US, 2020-2021.

During the COVID-19 pandemic, forecasting COVID-19 trends to support planning and response was a priority for scientists and decision makers alike. In the United States, COVID-19 forecasting was coordinated by a large group of universities, companies, and government entities led by the Centers for Disease Control and Prevention and the US COVID-19 Forecast Hub (https://covid19forecasthub.org). We evaluated approximately 9.7 million forecasts of weekly state-level COVID-19 cases for predictions 1-4 weeks into the future submitted by 24 teams from August 2020 to December 2021. We assessed coverage of central prediction intervals and weighted interval scores (WIS), adjusting for missing forecasts relative to a baseline forecast, and used a Gaussian generalized estimating equation (GEE) model to evaluate differences in skill across epidemic phases that were defined by the effective reproduction number. Overall, we found high variation in skill across individual models, with ensemble-based forecasts outperforming other approaches. Forecast skill relative to the baseline was generally higher for larger jurisdictions (e.g., states compared to counties). Over time, forecasts generally performed worst in periods of rapid changes in reported cases (either in increasing or decreasing epidemic phases) with 95% prediction interval coverage dropping below 50% during the growth phases of the winter 2020, Delta, and Omicron waves. Ideally, case forecasts could serve as a leading indicator of changes in transmission dynamics. However, while most COVID-19 case forecasts outperformed a naïve baseline model, even the most accurate case forecasts were unreliable in key phases. Further research could improve forecasts of leading indicators, like COVID-19 cases, by leveraging additional real-time data, addressing performance across phases, improving the characterization of forecast confidence, and ensuring that forecasts were coherent across spatial scales. In the meantime, it is critical for forecast users to appreciate current limitations and use a broad set of indicators to inform pandemic-related decision making.

Machine learning to predict the source of campylobacteriosis using whole genome data.

Campylobacteriosis is among the world's most common foodborne illnesses, caused predominantly by the bacterium Campylobacter jejuni. Effective interventions require determination of the infection source which is challenging as transmission occurs via multiple sources such as contaminated meat, poultry, and drinking water. Strain variation has allowed source tracking based upon allelic variation in multi-locus sequence typing (MLST) genes allowing isolates from infected individuals to be attributed to specific animal or environmental reservoirs. However, the accuracy of probabilistic attribution models has been limited by the ability to differentiate isolates based upon just 7 MLST genes. Here, we broaden the input data spectrum to include core genome MLST (cgMLST) and whole genome sequences (WGS), and implement multiple machine learning algorithms, allowing more accurate source attribution. We increase attribution accuracy from 64% using the standard iSource population genetic approach to 71% for MLST, 85% for cgMLST and 78% for kmerized WGS data using the classifier we named aiSource. To gain insight beyond the source model prediction, we use Bayesian inference to analyse the relative affinity of C. jejuni strains to infect humans and identified potential differences, in source-human transmission ability among clonally related isolates in the most common disease causing lineage (ST-21 clonal complex). Providing generalizable computationally efficient methods, based upon machine learning and population genetics, we provide a scalable approach to global disease surveillance that can continuously incorporate novel samples for source attribution and identify fine-scale variation in transmission potential.

Genome-wide association studies reveal the role of polymorphisms affecting factor H binding protein expression in host invasion by Neisseria meningitidis.

Many invasive bacterial diseases are caused by organisms that are ordinarily harmless components of the human microbiome. Effective interventions against these microbes require an understanding of the processes whereby symbiotic or commensal relationships transition into pathology. Here, we describe bacterial genome-wide association studies (GWAS) of Neisseria meningitidis, a common commensal of the human respiratory tract that is nevertheless a leading cause of meningitis and sepsis. An initial GWAS discovered bacterial genetic variants, including single nucleotide polymorphisms (SNPs), associated with invasive meningococcal disease (IMD) versus carriage in several loci across the meningococcal genome, encoding antigens and other extracellular components, confirming the polygenic nature of the invasive phenotype. In particular, there was a significant peak of association around the fHbp locus, encoding factor H binding protein (fHbp), which promotes bacterial immune evasion of human complement by recruiting complement factor H (CFH) to the meningococcal surface. The association around fHbp with IMD was confirmed by a validation GWAS, and we found that the SNPs identified in the validation affected the 5' region of fHbp mRNA, altering secondary RNA structures, thereby increasing fHbp expression and enhancing bacterial escape from complement-mediated killing. This finding is consistent with the known link between complement deficiencies and CFH variation with human susceptibility to IMD. These observations demonstrate the importance of human and bacterial genetic variation across the fHbp:CFH interface in determining IMD susceptibility, the transition from carriage to disease.

Hyperkalemia Risk with Finerenone: Results from the FIDELIO-DKD Trial.

BACKGROUND: Finerenone reduced risk of cardiorenal outcomes in patients with CKD and type 2 diabetes in the FIDELIO-DKD trial. We report incidences and risk factors for hyperkalemia with finerenone and placebo in FIDELIO-DKD. METHODS: This post hoc safety analysis defined hyperkalemia as ≥mild or ≥moderate based on serum potassium concentrations of >5.5 or >6.0 mmol/L, respectively, assessed at all regular visits. Cumulative incidences of hyperkalemia were based on the Aalen-Johansen estimator using death as competing risk. A multivariate Cox proportional hazards model identified significant independent predictors of hyperkalemia. Restricted cubic splines assessed relationships between short-term post-baseline changes in serum potassium or eGFR and subsequent hyperkalemia risk. During the study, serum potassium levels guided drug dosing. Patients in either group who experienced ≥mild hyperkalemia had the study drug withheld until serum potassium was ≤5.0 mmol/L; then the drug was restarted at the 10 mg daily dose. Placebo-treated patients underwent sham treatment interruption and downtitration. RESULTS: Over 2.6 years' median follow-up, 597 of 2785 (21.4%) and 256 of 2775 (9.2%) patients treated with finerenone and placebo, respectively, experienced treatment-emergent ≥mild hyperkalemia; 126 of 2802 (4.5%) and 38 of 2796 (1.4%) patients, respectively, experienced moderate hyperkalemia. Independent risk factors for ≥mild hyperkalemia were higher serum potassium, lower eGFR, increased urine albumin-creatinine ratio, younger age, female sex, ß-blocker use, and finerenone assignment. Diuretic or sodium-glucose cotransporter-2 inhibitor use reduced risk. In both groups, short-term increases in serum potassium and decreases in eGFR were associated with subsequent hyperkalemia. At month 4, the magnitude of increased hyperkalemia risk for any change from baseline was smaller with finerenone than with placebo. CONCLUSIONS: Finerenone was independently associated with hyperkalemia. However, routine potassium monitoring and hyperkalemia management strategies employed in FIDELIO-DKD minimized the impact of hyperkalemia, providing a basis for clinical use of finerenone.

5-((3-Amidobenzyl)oxy)nicotinamides as SIRT2 Inhibitors: A Study of Constrained Analogs.

SIRT2 is a member of NAD+-dependent sirtuins and its inhibition has been proposed as a promising therapeutic approach for treating human diseases, including neurodegenerative diseases, cancer, and infections. Expanding SIRT2 inhibitors based on the 3-aminobenzyloxy nicotinamide core structure, we have synthesized and evaluated constrained analogs and selected stereoisomers. Our structure-activity relationship (SAR) study has revealed that 2,3-constrained (S)-isomers possess enhanced in vitro enzymatic inhibitory activity against SIRT2 and retain excellent selectivity over SIRT1 and SIRT3, provided that a suitable ring A is used. This current study further explores SIRT2 inhibitors based on the 3-aminobenzyloxy nicotinamide scaffold and contributes to the discovery of potent, selective SIRT2 inhibitors that have been actively pursued for their potential therapeutic applications.

Evaluation of Monoexponential, Stretched-Exponential and Intravoxel Incoherent Motion MRI Diffusion Models in Early Response Monitoring to Neoadjuvant Chemotherapy in Patients With Breast Cancer-A Preliminary Study.

BACKGROUND: There has been a growing interest in exploring the applications of stretched-exponential (SEM) and intravoxel incoherent motion (IVIM) models of diffusion-weighted imaging (DWI) in breast imaging, with the focus on differentiation of breast lesions. However, the use of SEM and IVIM models to predict early response to neoadjuvant chemotherapy (NACT) has received less attention. PURPOSE: To investigate the value of monoexponential, SEM, and IVIM models to predict early response to NACT in patients with primary breast cancer. STUDY TYPE: Prospective. POPULATION: Thirty-seven patients with primary breast cancer (aged 46 ± 11 years) due to undergo NACT. FIELD STRENGTH/SEQUENCES: A 1.5-T MR scanner, T1 -weighted three-dimensional spoiled gradient-echo, two-dimensional single-shot spin-echo echo-planar imaging sequence (DWI) at six b-values (0-800 s mm-2 ). ASSESSMENT: Tumor volume, apparent diffusion coefficient, tissue diffusion (Dt ), pseudo-diffusion coefficient (Dp ), perfusion fraction (f), distributed diffusion coefficient, and alpha (α) were extracted, following volumetric sampling of the tumors, at three time-points: pretreatment, post one and three cycles of NACT. STATISTICAL TESTS: Mann-Whitney test, receiver operating characteristic (ROC) curve. Statistical significance level was P < 0.05. RESULTS: Following NACT, 17 patients were determined to be pathological responders and 20 nonresponders. Tumor volume was significantly larger in nonresponders at each MRI time-point and demonstrated reasonable performance in predicting response (area under the ROC curve [AUC] = 0.83-0.87). No significant differences between groups were found in the diffusion coefficients at each time-point (P = 0.09-1). The parameters α (SEM), f, and f × Dp (IVIM) were able to differentiate between response groups after one cycle of NACT (AUC = 0.73, 0.72, and 0.74, respectively). CONCLUSION: Diffusion coefficients derived from the monoexponential, SEM, and IVIM models did not predict pathological response. However, the IVIM-derived parameters f and f × Dp and the SEM-derived parameter α were able to predict response to NACT in breast cancer patients following one cycle of NACT. LEVEL OF EVIDENCE: 2 TECHNICAL EFFICACY STAGE: 2.

Bias, Repeatability and Reproducibility of Liver T1 Mapping With Variable Flip Angles.

BACKGROUND: Three-dimensional variable flip angle (VFA) methods are commonly used for T1 mapping of the liver, but there is no data on the accuracy, repeatability, and reproducibility of this technique in this organ in a multivendor setting. PURPOSE: To measure bias, repeatability, and reproducibility of VFA T1 mapping in the liver. STUDY TYPE: Prospective observational. POPULATION: Eight healthy volunteers, four women, with no known liver disease. FIELD STRENGTH/SEQUENCE: 1.5-T and 3.0-T; three-dimensional steady-state spoiled gradient echo with VFAs; Look-Locker. ASSESSMENT: Traveling volunteers were scanned twice each (30 minutes to 3 months apart) on six MRI scanners from three vendors (GE Healthcare, Philips Medical Systems, and Siemens Healthineers) at two field strengths. The maximum period between the first and last scans among all volunteers was 9 months. Volunteers were instructed to abstain from alcohol intake for at least 72 hours prior to each scan and avoid high cholesterol foods on the day of the scan. STATISTICAL TESTS: Repeated measures ANOVA, Student t-test, Levene's test of variances, and 95% significance level. The percent error relative to literature liver T1 in healthy volunteers was used to assess bias. The relative error (RE) due to intrascanner and interscanner variation in T1 measurements was used to assess repeatability and reproducibility. RESULTS: The 95% confidence interval (CI) on the mean bias and mean repeatability RE of VFA T1 in the healthy liver was 34 ± 6% and 10 ± 3%, respectively. The 95% CI on the mean reproducibility RE at 1.5 T and 3.0 T was 29 ± 7% and 25 ± 4%, respectively. DATA CONCLUSION: Bias, repeatability, and reproducibility of VFA T1 mapping in the liver in a multivendor setting are similar to those reported for breast, prostate, and brain. LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY STAGE: 1.

Probabilistic transmission models incorporating sequencing data for healthcare-associated Clostridioides difficile outperform heuristic rules and identify strain-specific differences in transmission.

Fitting stochastic transmission models to electronic patient data can offer detailed insights into the transmission of healthcare-associated infections and improve infection control. Pathogen whole-genome sequencing may improve the precision of model inferences, but computational constraints have limited modelling applications predominantly to small datasets and specific outbreaks, whereas large-scale sequencing studies have mostly relied on simple rules for identifying/excluding plausible transmission. We present a novel approach for integrating detailed epidemiological data on patient contact networks in hospitals with large-scale pathogen sequencing data. We apply our approach to study Clostridioides difficile transmission using a dataset of 1223 infections in Oxfordshire, UK, 2007-2011. 262 (21% [95% credibility interval 20-22%]) infections were estimated to have been acquired from another known case. There was heterogeneity by sequence type (ST) in the proportion of cases acquired from another case with the highest rates in ST1 (ribotype-027), ST42 (ribotype-106) and ST3 (ribotype-001). These same STs also had higher rates of transmission mediated via environmental contamination/spores persisting after patient discharge/recovery; for ST1 these persisted longer than for most other STs except ST3 and ST42. We also identified variation in transmission between hospitals, medical specialties and over time; by 2011 nearly all transmission from known cases had ceased in our hospitals. Our findings support previous work suggesting only a minority of C. difficile infections are acquired from known cases but highlight a greater role for environmental contamination than previously thought. Our approach is applicable to other healthcare-associated infections. Our findings have important implications for effective control of C. difficile.

The harmonic mean p-value for combining dependent tests.

Analysis of "big data" frequently involves statistical comparison of millions of competing hypotheses to discover hidden processes underlying observed patterns of data, for example, in the search for genetic determinants of disease in genome-wide association studies (GWAS). Controlling the familywise error rate (FWER) is considered the strongest protection against false positives but makes it difficult to reach the multiple testing-corrected significance threshold. Here, I introduce the harmonic mean p-value (HMP), which controls the FWER while greatly improving statistical power by combining dependent tests using generalized central limit theorem. I show that the HMP effortlessly combines information to detect statistically significant signals among groups of individually nonsignificant hypotheses in examples of a human GWAS for neuroticism and a joint human-pathogen GWAS for hepatitis C viral load. The HMP simultaneously tests all ways to group hypotheses, allowing the smallest groups of hypotheses that retain significance to be sought. The power of the HMP to detect significant hypothesis groups is greater than the power of the Benjamini-Hochberg procedure to detect significant hypotheses, although the latter only controls the weaker false discovery rate (FDR). The HMP has broad implications for the analysis of large datasets, because it enhances the potential for scientific discovery.

GenomegaMap: Within-Species Genome-Wide dN/dS Estimation from over 10,000 Genomes.

The dN/dS ratio provides evidence of adaptation or functional constraint in protein-coding genes by quantifying the relative excess or deficit of amino acid-replacing versus silent nucleotide variation. Inexpensive sequencing promises a better understanding of parameters, such as dN/dS, but analyzing very large data sets poses a major statistical challenge. Here, I introduce genomegaMap for estimating within-species genome-wide variation in dN/dS, and I apply it to 3,979 genes across 10,209 tuberculosis genomes to characterize the selection pressures shaping this global pathogen. GenomegaMap is a phylogeny-free method that addresses two major problems with existing approaches: 1) It is fast no matter how large the sample size and 2) it is robust to recombination, which causes phylogenetic methods to report artefactual signals of adaptation. GenomegaMap uses population genetics theory to approximate the distribution of allele frequencies under general, parent-dependent mutation models. Coalescent simulations show that substitution parameters are well estimated even when genomegaMap's simplifying assumption of independence among sites is violated. I demonstrate the ability of genomegaMap to detect genuine signatures of selection at antimicrobial resistance-conferring substitutions in Mycobacterium tuberculosis and describe a novel signature of selection in the cold-shock DEAD-box protein A gene deaD/csdA. The genomegaMap approach helps accelerate the exploitation of big data for gaining new insights into evolution within species.

4,7-Disubstituted 7H-Pyrrolo[2,3-d]pyrimidines and Their Analogs as Antiviral Agents against Zika Virus.

Discovery of compound 1 as a Zika virus (ZIKV) inhibitor has prompted us to investigate its 7H-pyrrolo[2,3-d]pyrimidine scaffold, revealing structural features that elicit antiviral activity. Furthermore, we have demonstrated that 9H-purine or 1H-pyrazolo[3,4-d]pyrimidine can serve as an alternative core structure. Overall, we have identified 4,7-disubstituted 7H-pyrrolo[2,3-d]pyrimidines and their analogs including compounds 1, 8 and 11 as promising antiviral agents against flaviviruses ZIKV and dengue virus (DENV). While the molecular target of these compounds is yet to be elucidated, 4,7-disubstituted 7H-pyrrolo[2,3-d]pyrimidines and their analogs are new chemotypes in the design of small molecules against flaviviruses, an important group of human pathogens.

Application of machine learning techniques to tuberculosis drug resistance analysis.

MOTIVATION: Timely identification of Mycobacterium tuberculosis (MTB) resistance to existing drugs is vital to decrease mortality and prevent the amplification of existing antibiotic resistance. Machine learning methods have been widely applied for timely predicting resistance of MTB given a specific drug and identifying resistance markers. However, they have been not validated on a large cohort of MTB samples from multi-centers across the world in terms of resistance prediction and resistance marker identification. Several machine learning classifiers and linear dimension reduction techniques were developed and compared for a cohort of 13 402 isolates collected from 16 countries across 6 continents and tested 11 drugs. RESULTS: Compared to conventional molecular diagnostic test, area under curve of the best machine learning classifier increased for all drugs especially by 23.11%, 15.22% and 10.14% for pyrazinamide, ciprofloxacin and ofloxacin, respectively (P < 0.01). Logistic regression and gradient tree boosting found to perform better than other techniques. Moreover, logistic regression/gradient tree boosting with a sparse principal component analysis/non-negative matrix factorization step compared with the classifier alone enhanced the best performance in terms of F1-score by 12.54%, 4.61%, 7.45% and 9.58% for amikacin, moxifloxacin, ofloxacin and capreomycin, respectively, as well increasing area under curve for amikacin and capreomycin. Results provided a comprehensive comparison of various techniques and confirmed the application of machine learning for better prediction of the large diverse tuberculosis data. Furthermore, mutation ranking showed the possibility of finding new resistance/susceptible markers. AVAILABILITY AND IMPLEMENTATION: The source code can be found at http://www.robots.ox.ac.uk/ davidc/code.php. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

DeepAMR for predicting co-occurrent resistance of Mycobacterium tuberculosis.

MOTIVATION: Resistance co-occurrence within first-line anti-tuberculosis (TB) drugs is a common phenomenon. Existing methods based on genetic data analysis of Mycobacterium tuberculosis (MTB) have been able to predict resistance of MTB to individual drugs, but have not considered the resistance co-occurrence and cannot capture latent structure of genomic data that corresponds to lineages. RESULTS: We used a large cohort of TB patients from 16 countries across six continents where whole-genome sequences for each isolate and associated phenotype to anti-TB drugs were obtained using drug susceptibility testing recommended by the World Health Organization. We then proposed an end-to-end multi-task model with deep denoising auto-encoder (DeepAMR) for multiple drug classification and developed DeepAMR_cluster, a clustering variant based on DeepAMR, for learning clusters in latent space of the data. The results showed that DeepAMR outperformed baseline model and four machine learning models with mean AUROC from 94.4% to 98.7% for predicting resistance to four first-line drugs [i.e. isoniazid (INH), ethambutol (EMB), rifampicin (RIF), pyrazinamide (PZA)], multi-drug resistant TB (MDR-TB) and pan-susceptible TB (PANS-TB: MTB that is susceptible to all four first-line anti-TB drugs). In the case of INH, EMB, PZA and MDR-TB, DeepAMR achieved its best mean sensitivity of 94.3%, 91.5%, 87.3% and 96.3%, respectively. While in the case of RIF and PANS-TB, it generated 94.2% and 92.2% sensitivity, which were lower than baseline model by 0.7% and 1.9%, respectively. t-SNE visualization shows that DeepAMR_cluster captures lineage-related clusters in the latent space. AVAILABILITY AND IMPLEMENTATION: The details of source code are provided at http://www.robots.ox.ac.uk/∼davidc/code.php. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Bayesian inference of ancestral dates on bacterial phylogenetic trees.

The sequencing and comparative analysis of a collection of bacterial genomes from a single species or lineage of interest can lead to key insights into its evolution, ecology or epidemiology. The tool of choice for such a study is often to build a phylogenetic tree, and more specifically when possible a dated phylogeny, in which the dates of all common ancestors are estimated. Here, we propose a new Bayesian methodology to construct dated phylogenies which is specifically designed for bacterial genomics. Unlike previous Bayesian methods aimed at building dated phylogenies, we consider that the phylogenetic relationships between the genomes have been previously evaluated using a standard phylogenetic method, which makes our methodology much faster and scalable. This two-step approach also allows us to directly exploit existing phylogenetic methods that detect bacterial recombination, and therefore to account for the effect of recombination in the construction of a dated phylogeny. We analysed many simulated datasets in order to benchmark the performance of our approach in a wide range of situations. Furthermore, we present applications to three different real datasets from recent bacterial genomic studies. Our methodology is implemented in a R package called BactDating which is freely available for download at https://github.com/xavierdidelot/BactDating.

Epidemics And Disability.

This essay argues that considering disability and disability history needs to be part of any history of epidemics. Recent scholarship has shown the many intersections of disability history and history of medicine. This essay argues that disability plays many roles in an epidemic from establishing pre-existing conditions, to affecting the acute phase of the disease, to creating lingering disabilities in the long aftermath. Histories of epidemics that ignore the many ways in which disability affects the experience of an epidemic are incomplete.

Machine learning for classifying tuberculosis drug-resistance from DNA sequencing data.

Motivation: Correct and rapid determination of Mycobacterium tuberculosis (MTB) resistance against available tuberculosis (TB) drugs is essential for the control and management of TB. Conventional molecular diagnostic test assumes that the presence of any well-studied single nucleotide polymorphisms is sufficient to cause resistance, which yields low sensitivity for resistance classification. Summary: Given the availability of DNA sequencing data from MTB, we developed machine learning models for a cohort of 1839 UK bacterial isolates to classify MTB resistance against eight anti-TB drugs (isoniazid, rifampicin, ethambutol, pyrazinamide, ciprofloxacin, moxifloxacin, ofloxacin, streptomycin) and to classify multi-drug resistance. Results: Compared to previous rules-based approach, the sensitivities from the best-performing models increased by 2-4% for isoniazid, rifampicin and ethambutol to 97% (P < 0.01), respectively; for ciprofloxacin and multi-drug resistant TB, they increased to 96%. For moxifloxacin and ofloxacin, sensitivities increased by 12 and 15% from 83 and 81% based on existing known resistance alleles to 95% and 96% (P < 0.01), respectively. Particularly, our models improved sensitivities compared to the previous rules-based approach by 15 and 24% to 84 and 87% for pyrazinamide and streptomycin (P < 0.01), respectively. The best-performing models increase the area-under-the-ROC curve by 10% for pyrazinamide and streptomycin (P < 0.01), and 4-8% for other drugs (P < 0.01). Availability and implementation: The details of source code are provided at http://www.robots.ox.ac.uk/~davidc/code.php. Contact: david.clifton@eng.ox.ac.uk. Supplementary information: Supplementary data are available at Bioinformatics online.

A functional form for a representative individual arterial input function measured from a population using high temporal resolution DCE MRI.

PURPOSE: To measure the arterial input function (AIF), an essential component of tracer kinetic analysis, in a population of patients using an optimized dynamic contrast-enhanced (DCE) imaging sequence and to estimate inter- and intrapatient variability. From these data, a representative AIF that may be used for realistic simulation studies can be extracted. METHODS: Thirty-nine female patients were imaged on multiple visits before and during a course of neoadjuvant chemotherapy for breast cancer. A total of 97 T1 -weighted DCE studies were analyzed including bookend estimates of T1 and model-fitting to each individual AIF. Area under the curve and cardiac output were estimated from each first pass peak, and these data were used to assess inter- and intrapatient variability of the AIF. RESULTS: Interpatient variability exceeded intrapatient variability of the AIF. There was no change in cardiac output as a function of MR visit (mean value 5.6 ± 1.1 L/min) but baseline blood T1 increased significantly following the start of chemotherapy (which was accompanied by a decrease in hematocrit). CONCLUSION: The AIF in an individual patient can be measured reproducibly but the variability of AIFs between patients suggests that use of a population AIF will decrease the precision of tracer kinetic analysis performed in cross-patient comparison studies. A representative AIF is presented that is typical of the population but retains the characteristics of an individually measured AIF.

Bayesian reconstruction of transmission within outbreaks using genomic variants.

Pathogen genome sequencing can reveal details of transmission histories and is a powerful tool in the fight against infectious disease. In particular, within-host pathogen genomic variants identified through heterozygous nucleotide base calls are a potential source of information to identify linked cases and infer direction and time of transmission. However, using such data effectively to model disease transmission presents a number of challenges, including differentiating genuine variants from those observed due to sequencing error, as well as the specification of a realistic model for within-host pathogen population dynamics. Here we propose a new Bayesian approach to transmission inference, BadTrIP (BAyesian epiDemiological TRansmission Inference from Polymorphisms), that explicitly models evolution of pathogen populations in an outbreak, transmission (including transmission bottlenecks), and sequencing error. BadTrIP enables the inference of host-to-host transmission from pathogen sequencing data and epidemiological data. By assuming that genomic variants are unlinked, our method does not require the computationally intensive and unreliable reconstruction of individual haplotypes. Using simulations we show that BadTrIP is robust in most scenarios and can accurately infer transmission events by efficiently combining information from genetic and epidemiological sources; thanks to its realistic model of pathogen evolution and the inclusion of epidemiological data, BadTrIP is also more accurate than existing approaches. BadTrIP is distributed as an open source package (https://bitbucket.org/nicofmay/badtrip) for the phylogenetic software BEAST2. We apply our method to reconstruct transmission history at the early stages of the 2014 Ebola outbreak, showcasing the power of within-host genomic variants to reconstruct transmission events.

Natural mutations in a Staphylococcus aureus virulence regulator attenuate cytotoxicity but permit bacteremia and abscess formation.

Staphylococcus aureus is a major bacterial pathogen, which causes severe blood and tissue infections that frequently emerge by autoinfection with asymptomatically carried nose and skin populations. However, recent studies report that bloodstream isolates differ systematically from those found in the nose and skin, exhibiting reduced toxicity toward leukocytes. In two patients, an attenuated toxicity bloodstream infection evolved from an asymptomatically carried high-toxicity nasal strain by loss-of-function mutations in the gene encoding the transcription factor repressor of surface proteins (rsp). Here, we report that rsp knockout mutants lead to global transcriptional and proteomic reprofiling, and they exhibit the greatest signal in a genome-wide screen for genes influencing S. aureus survival in human cells. This effect is likely to be mediated in part via SSR42, a long-noncoding RNA. We show that rsp controls SSR42 expression, is induced by hydrogen peroxide, and is required for normal cytotoxicity and hemolytic activity. Rsp inactivation in laboratory- and bacteremia-derived mutants attenuates toxin production, but up-regulates other immune subversion proteins and reduces lethality during experimental infection. Crucially, inactivation of rsp preserves bacterial dissemination, because it affects neither formation of deep abscesses in mice nor survival in human blood. Thus, we have identified a spontaneously evolving, attenuated-cytotoxicity, nonhemolytic S. aureus phenotype, controlled by a pleiotropic transcriptional regulator/noncoding RNA virulence regulatory system, capable of causing S. aureus bloodstream infections. Such a phenotype could promote deep infection with limited early clinical manifestations, raising concerns that bacterial evolution within the human body may contribute to severe infection.

The distribution of bacterial doubling times in the wild.

Generation time varies widely across organisms and is an important factor in the life cycle, life history and evolution of organisms. Although the doubling time (DT) has been estimated for many bacteria in the laboratory, it is nearly impossible to directly measure it in the natural environment. However, an estimate can be obtained by measuring the rate at which bacteria accumulate mutations per year in the wild and the rate at which they mutate per generation in the laboratory. If we assume the mutation rate per generation is the same in the wild and in the laboratory, and that all mutations in the wild are neutral, an assumption that we show is not very important, then an estimate of the DT can be obtained by dividing the latter by the former. We estimate the DT for five species of bacteria for which we have both an accumulation and a mutation rate estimate. We also infer the distribution of DTs across all bacteria from the distribution of the accumulation and mutation rates. Both analyses suggest that DTs for bacteria in the wild are substantially greater than those in the laboratory, that they vary by orders of magnitude between different species of bacteria and that a substantial fraction of bacteria double very slowly in the wild.