Chieh-Hsi Wu

The spread of infectious diseases is shaped by spatial and temporal aspects, such as host population structure or changes in the transmission rate or number of infected individuals over time. These spatiotemporal dynamics are imprinted in the genome of pathogens and can be recovered from those genomes using phylodynamics methods. However, phylodynamic methods typically quantify either the temporal or spatial transmission dynamics, which leads to unclear biases, as one can potentially not be inferred without the other. Here, we address this challenge by introducing a structured coalescent skyline approach, MASCOT-Skyline that allows us to jointly infer spatial and temporal transmission dynamics of infectious diseases using Markov chain Monte Carlo inference. To do so, we model the effective population size dynamics in different locations using a non-parametric function, allowing us to approximate a range of population size dynamics. We show, using a range of different viral outbreak datasets, potential issues with phylogeographic methods. We then use these viral datasets to motivate simulations of outbreaks that illuminate the nature of biases present in the different phylogeographic methods. We show that spatial and temporal dynamics should be modeled jointly even if one seeks to recover just one of the two. Further, we showcase conditions under which we can expect phylogeographic analyses to be biased, particularly different subsampling approaches, as well as provide recommendations of when we can expect them to perform well. We implemented MASCOT-Skyline as part of the open-source software package MASCOT for the Bayesian phylodynamics platform BEAST2.

Objective Dose shortages delayed access to COVID-19 vaccination. We aim to characterise inequality in two-dose vaccination by sociodemographic group across Brazil. Design This is a cross-sectional study. Setting We used data retrieved from the Brazilian Ministry of Health databases published between 17 January 2021 and 6 September 2021. Methods We assessed geographical inequalities in full vaccination coverage and dose by age, sex, race and socioeconomic status. We developed a Campaign Optimality Index to characterise inequality in vaccination access due to premature vaccination towards younger populations before older and vulnerable populations were fully vaccinated. Generalised linear regression was used to investigate the risk of death and hospitalisation by age group, socioeconomic status and vaccination coverage. Results Vaccination coverage is higher in the wealthier South and Southeast. Men, people of colour and low-income groups were more likely to be only partially vaccinated due to missing or delaying a second dose. Vaccination started prematurely for age groups under 50 years which may have hindered uptake in older age groups. Vaccination coverage was associated with a lower risk of death, especially in older age groups (ORs 9.7 to 29.0, 95% CI 9. 4 to 29.9). Risk of hospitalisation was greater in areas with higher vaccination rates due to higher access to care and reporting. Conclusions Vaccination inequality persists between states, age and demographic groups despite increasing uptake. The association between hospitalisation rates and vaccination is attributed to preferential delivery to areas of greater transmission and access to healthcare.

Classification of unlabeled data is usually achieved by supervised learning from labeled samples. Although there exist many sophisticated supervised machine learning methods that can predict the missing labels with a high level of accuracy, they often lack the required transparency in situations where it is important to provide interpretable results and meaningful measures of confidence. Body fluid classification of forensic casework data is the case in point. We develop a new Biclustering Dirichlet Process for Class-assignment with Random Matrices (BDP-CaRMa), with a three-level hierarchy of clustering, and a model-based approach to classification that adapts to block structure in the data matrix. As the class labels of some observations are missing, the number of rows in the data matrix for each class is unknown. BDP-CaRMa handles this and extends existing biclustering methods by simultaneously biclustering multiple matrices each having a randomly variable number of rows. We demonstrate our method by applying it to the motivating problem, which is the classification of body fluids based on mRNA profiles taken from crime scenes. The analyses of casework-like data show that our method is interpretable and produces well-calibrated posterior probabilities. Our model can be more generally applied to other types of data with a similar structure to the forensic data.

The COVID-19 pandemic has caused over half a million deaths in Brazil, and public healthcare nearly collapsed. Vaccination differs between states and demographics. Dose shortages delayed access. In this cross-sectional study, data were retrieved from the Brazilian Ministry of Health databases published since 17 January 2021, respectively. We developed a campaign optimality index to characterise inequality in vaccination access caused by age due to premature vaccination towards younger populations before older and vulnerable populations were fully vaccinated. We assessed geographical inequalities in full vaccination coverage and dose by age, sex, race, and socioeconomic status. Generalised linear regression was used to investigate the risk of death and hospitalisation by age group, socioeconomic status, and vaccination coverage. Vaccination coverage is higher in the wealthier South and Southeast. Men, people of colour, and low-income groups were more likely to be only partially vaccinated due to missing or delaying a second dose. Vaccination started prematurely for age groups under 50 years and may have hindered uptake of older age groups. Vaccination coverage was associated with a lower risk of death, especially in older age groups (OR: 10.5–34.8, 95% CI: (10.2, 35.9)). Risk of hospitalisation was greater in areas with higher vaccination rates due to higher access to care and reporting. Vaccination inequality persists between states, age and demographic groups despite increasing uptake. The association between hospitalisation rates and vaccination is attributed to preferential delivery to areas of greater transmission and access to healthcare.

Neurological complications such as stroke and encephalopathies affect 4% of adults with acute COVID-19 infection, causing significant morbidity with long term health and economic consequences. Dexamethasone reduces the 28-day mortality in hospitalised patients requiring supplemental oxygen, while remdesivir has antiviral action against SARS-CoV-2 and reduces the severity of COVID-19 in these patients. We hypothesise that adding either remdesivir, dexamethasone or both to standard of care in patients hospitalised with COVID-19 may reduce the incidence of COVID-19-associated neurological complications. This retrospective analysis of the ISARIC4C cohort will analyse data from patients aged 18 years and older admitted to hospital with COVID-19 between 31 Jan 2020 and 29 Jun 2021. The incidence of neurological complications following COVID-19 in patients receiving standard of care and either remdesivir, dexamethasone or both will be compared against the standard of care using propensity scoring methodology. The effect of neurological complications on markers of disease severity, as well as clinical outcome, will be assessed using multivariable linear and logistic regression.

Staphylococcus aureus is a major bacterial pathogen in humans, and a dominant cause of severe bloodstream infections. Globally, antimicrobial resistance (AMR) in S. aureus remains challenging. While human risk factors for infection have been defined, contradictory evidence exists for the role of bacterial genomic variation in S. aureus disease. To investigate the contribution of bacterial lineage and genomic variation to the development of bloodstream infection, we undertook a genome-wide association study comparing bacteria from 1017 individuals with bacteraemia to 984 adults with asymptomatic S. aureus nasal carriage. Within 984 carriage isolates, we also compared healthcare-associated (HA) carriage with community-associated (CA) carriage. All major global lineages were represented in both bacteraemia and carriage, with no evidence for different infection rates. However, kmers tagging trimethoprim resistance-conferring mutation F99Y in dfrB were significantly associated with bacteraemia-vs-carriage (P=10^-8.9-10^-9.3). Pooling variation within genes, bacteraemia-vs-carriage was associated with the presence of mecA (HMP=10^-5.3) as well as the presence of SCCmec (HMP=10^-4.4). Among S. aureus carriers, no lineages were associated with HA-vs-CA carriage. However, we found a novel signal of HA-vs-CA carriage in the foldase protein prsA, where kmers representing conserved sequence allele were associated with CA carriage (P=10^-7.1-10^-19.4), while in gyrA, a ciprofloxacin resistance-conferring mutation, L84S, was associated with HA carriage (P=10^-7.2). In an extensive study of S. aureus bacteraemia and nasal carriage in the UK, we found strong evidence that all S. aureus lineages are equally capable of causing bloodstream infection, and of being carried in the healthcare environment. Genomic variation in the foldase protein prsA is a novel genomic marker of healthcare origin in S. aureus but was not associated with bacteraemia. AMR determinants were associated with both bacteraemia and healthcare-associated carriage, suggesting that AMR increases the propensity not only to survive in healthcare environments, but also to cause invasive disease.

Unmitigated spread in Brazil: Despite an extensive network of primary care availability, Brazil has suffered profoundly during the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic. Using daily data from state health offices, Castro et al. analyzed the pattern of spread of COVID-19 cases and deaths in the country from February to October 2020. Clusters of deaths before cases became apparent indicated unmitigated spread. SARS-CoV-2 circulated undetected in Brazil for more than a month as it spread north from Sã o Paulo. In Manaus, transmission reached unprecedented levels after a momentary respite in mid-2020. Faria et al. tracked the evolution of a new, more aggressive lineage called P.1, which has 17 mutations, including three (K417T, E484K, and N501Y) in the spike protein. After a period of accelerated evolution, this variant emerged in Brazil during November 2020. Coupled with the emergence of P.1, disease spread was accelerated by stark local inequalities and political upheaval, which compromised a prompt federal response. Science , abh1558 and abh2644, this issue p. 821 and p. 815

Introduction: Little evidence exists on the differential health effects of COVID-19 on disadvantaged population groups. Here we characterise the differential risk of hospitalisation and death in São Paulo state, Brazil, and show how vulnerability to COVID-19 is shaped by socioeconomic inequalities. Methods: We conducted a cross-sectional study using hospitalised severe acute respiratory infections notified from March to August 2020 in the Sistema de Monitoramento Inteligente de São Paulo database. We examined the risk of hospitalisation and death by race and socioeconomic status using multiple data sets for individual-level and spatiotemporal analyses. We explained these inequalities according to differences in daily mobility from mobile phone data, teleworking behaviour and comorbidities. Results: Throughout the study period, patients living in the 40% poorest areas were more likely to die when compared with patients living in the 5% wealthiest areas (OR: 1.60, 95% CI 1.48 to 1.74) and were more likely to be hospitalised between April and July 2020 (OR: 1.08, 95% CI 1.04 to 1.12). Black and Pardo individuals were more likely to be hospitalised when compared with White individuals (OR: 1.41, 95% CI 1.37 to 1.46; OR: 1.26, 95% CI 1.23 to 1.28, respectively), and were more likely to die (OR: 1.13, 95% CI 1.07 to 1.19; 1.07, 95% CI 1.04 to 1.10, respectively) between April and July 2020. Once hospitalised, patients treated in public hospitals were more likely to die than patients in private hospitals (OR: 1.40%, 95% CI 1.34% to 1.46%). Black individuals and those with low education attainment were more likely to have one or more comorbidities, respectively (OR: 1.29, 95% CI 1.19 to 1.39; 1.36, 95% CI 1.27 to 1.45). Conclusions: Low-income and Black and Pardo communities are more likely to die with COVID-19. This is associated with differential access to quality healthcare, ability to self-isolate and the higher prevalence of comorbidities.

Background: Staphylococcus aureus is a major bacterial pathogen in humans, and a dominant cause of severe bloodstream infections. Globally, antimicrobial resistance (AMR) in S. aureus remains challenging. While human risk factors for infection have been defined, contradictory evidence exists for the role of bacterial genomic variation in S. aureus disease. Methods: To investigate the contribution of bacterial lineage and genomic variation to the development of bloodstream infection, we undertook a genome-wide association study comparing bacteria from 1017 individuals with bacteraemia to 984 adults with asymptomatic S. aureus nasal carriage. Within 984 carriage isolates, we also compared healthcare-associated (HA) carriage with community-associated (CA) carriage. Results: All major global lineages were represented in both bacteraemia and carriage, with no evidence for different attack rates. However, kmers tagging trimethoprim resistance-conferring mutation F99Y in dfrB were significantly associated with bacteraemia-vs-carriage ( p= 10 ^−8.9 -10 ^−9.3 ). Pooling variation within genes, bacteraemia-vs-carriage was associated with the presence of mecA (HMP=10 ^−5.3 ) as well as the presence of SCCmec (HMP=10 ^−4.4 ). Among S. aureus carriers, no lineages were associated with HA-vs-CA carriage. However, we found a novel signal of HA-vs-CA carriage in the foldase protein prsA , where kmers representing conserved sequence allele were associated with CA carriage ( p= 10 ^−7.1 -10 ^−19.4 ), while in gyrA , a ciprofloxacin resistance-conferring mutation, L84S, was associated with HA carriage ( p= 10 ^−7.2 ). Conclusions: In an extensive study of S. aureus bacteraemia and nasal carriage in the UK, we found strong evidence that all S. aureus lineages are equally capable of causing bloodstream infection, and of being carried in the healthcare environment. Genomic variation in the foldase protein prsA is a novel genomic marker of healthcare origin in S. aureus but was not associated with bacteraemia. AMR determinants were associated with both bacteraemia and hospital-associated carriage, suggesting that AMR increases the propensity not only to survive in hospital environments, but also to cause invasive disease.