iiinova — where your digital self keeps evolving

David Simons @davidsimons.bsky.social · 24/06/2026

We should validate this with serosurveys stratified across the built-up gradient in West African cities. Crucially, if we want to power Phase III Lassa vaccine trials, we need to target peri-urban 'silent districts'. Locations where transmission may be ubiquitous but clinical reporting is absent.

000

David Simons @davidsimons.bsky.social · 24/06/2026

The scale of the burden is substantial, accounting for 'shielding' and antibody waning yields an estimated 2.6m annual LASV infections. Given this does not match reported clinical cases or mortality, the epidemiological missing is huge. A vast proportion of these infections must be paucisymptomatic.

100

David Simons @davidsimons.bsky.social · 24/06/2026

To reproduce this, I built an Integrated Multi-Species Occupancy Model for the host community, using rodent surveillance to map the viral hazard. Calibrating this against human seroprevalence allowed me to estimate annual infection whilst modelling the protective 'shield' of the built environment.

100

David Simons @davidsimons.bsky.social · 04/06/2026

7/7 The punchline: pristine, hyper-diverse tropical forests actually exhibit the lowest mean intrinsic hazard. Anthropogenic disturbance acts as an ecological filter, stripping away slow-lived specialists and leaving a high-risk assembly of fast-lived, zoonotic amplifiers.

A global bivariate choropleth map illustrating the intersection of small mammal species richness and mean intrinsic zoonotic hazard (community competence). A three-by-three colour matrix serves as the legend. Teal regions, such as the Amazon basin and Southeast Asia, indicate high species richness but low mean hazard. Pink regions, including northern latitudes and arid zones like Australia, indicate low species richness but high mean hazard. Dark purple regions, such as West and East Africa, Central America, and parts of South America, represent overlapping hotspots of both high biodiversity and high intrinsic hazard. Areas with fewer than two modelled species, including Antarctica and Greenland, are masked in dark grey.

1181

David Simons @davidsimons.bsky.social · 04/06/2026

6/7 Are these host-virus links ancient? Evolutionary models show broad co-divergence but frequent host-switching. Phylogenetic accumulation curves reveal these novel hosts are almost entirely discovered in reactive bursts immediately following major human outbreaks.

Four panels detailing host phylogenetic diversity (PD) discovery over time. Panel A is a scatter plot of annual change in host PD (ΔPD_t) for arenaviruses. Each point represents a host-year discovery event, coloured by calendar year, with outlined circles indicating the yearly maximum. Panel B shows the equivalent annual change plot for hantaviruses. Spearman trend statistics for yearly maxima and all points are provided in both panels. Panel C displays a line graph of cumulative host PD for arenaviruses based on empirical discovery order. A coloured line tracks the observed trajectory against a red logarithmic least-squares fit and a hatched band representing the 95% null envelope from 10,000 random permutations. Panel D mirrors Panel C for hantaviruses. Light grey vertical bands in Panels C and D mark successive discovery-year intervals.

151

David Simons @davidsimons.bsky.social · 04/06/2026

5/7 A fast pace of life (early maturity, massive litters) and synanthropy independently predict reservoir status. Those obligate commensal rodents thriving in human-modified landscapes are associated with viral infection, likely through tolerating and maintaining these viruses.

Three panels illustrating Bayesian phylogenetic model outputs for reservoir status. Panel A shows posterior distributions for pace of life and sampling effort coefficients. A slow pace of life is negatively associated with reservoir status, whereas increased sampling effort is strongly positively associated. Panel B is a line graph showing the marginal effect of surveillance bias, demonstrating a substantially elevated probability of reservoir detection as the number of sampled individuals increases. Panel C is a line graph showing that the probability of reservoir status strictly decreases as the pace of life transitions from fast to slow.

1111

David Simons @davidsimons.bsky.social · 04/06/2026

3/7 The scale of the anthropogenic filter was surprising. After compiling records for ~590,000 small mammals, we found 46% of rodent and eulipotyphlan genera remain unsampled. Global sampling correlates strictly with night lights and road access, not ecological relevance.

Three bar charts detailing sampling biases in small mammals. Panel A shows taxonomic sampling coverage for rodents, shrews, and hedgehogs; no single genus has more than 50% of its constituent species sampled. Panel B displays sampling bias by synanthropy state: non-synanthropic (NS), occasionally synanthropic (OS), and totally synanthropic (TS). The majority of species are categorised as unknown, but among known states, sampling effort increases progressively from NS to TS. Panel C shows the number of unique viruses tested per species. The distribution is heavily right-skewed; most are tested for 0 to 1 virus, decaying towards 10, with a small secondary peak at 10 and a long tail indicating a few highly investigated species tested for over 30 viruses.

Three line graphs displaying marginal effect plots for spatial surveillance bias predictors. Panel B shows a positive, increasing trend in sampling effort correlated with higher night-time light intensity. Panel C shows a negative trend, with sampling effort decreasing as remoteness (measured by travel time from major cities) increases. Panel D demonstrates that local host species richness has a flat, negligible effect on sampling effort.

180

David Simons @davidsimons.bsky.social · 04/06/2026

A global bivariate map illustrating the intersection of small mammal species richness and mean intrinsic zoonotic hazard (community competence). A 3x3 colour matrix serves as the legend. Teal regions, such as the Amazon basin and Southeast Asia, indicate high species richness but low mean hazard. Pink regions, including northern latitudes and arid zones like Australia, indicate low species richness but high mean hazard. Dark purple regions, such as West and East Africa, Central America and parts of South America, represent overlapping hotspots of both high biodiversity and high intrinsic hazard. Areas with fewer than two modelled species, such as Antarctica and Greenland, are masked in dark grey.

100

David Simons @davidsimons.bsky.social · 04/06/2026

a) Annual change in host phylogenetic diversity (ΔPD_t) for arenaviruses. Each point is a host-year discovery event, coloured by calendar year; outlined circles indicate the yearly maximum ΔPD_t event. b) Annual ΔPD_t for hantaviruses, with the same point encoding and yearly maximum outlines. Spearman trend statistics for yearly maxima (r_max) and all points (r_all) are shown in a) and b) (one-tailed p-values). c) Cumulative host PD for arenaviruses as hosts are added in empirical discovery order. The hatched band is the 95% null envelope from 10,000 random permutations of host discovery order; the observed trajectory is shown as the coloured line and points, and the red curve is the logarithmic least-squares fit. d) Cumulative host PD for hantaviruses, shown as in c). Light grey vertical bands in c) and d) mark successive discovery-year intervals.

100

David Simons @davidsimons.bsky.social · 04/06/2026

Figure a) shows the posterior distributions of the PGLMM for pace of life and sampling effort. A slow pace of life is negatively associated with being a reservoir, while increased sampling effort is associated with a substantially elevated probability of being a reservoir. b) shows the marginal effect of surveillance bias alone with an increasing probability of detection with increasing number of individuals sampled. c) shows that as pace of life moves from "fast" to "slow" the probability of reservoir status decreases.

100

David Simons @davidsimons.bsky.social · 04/06/2026

Bar chart A) shows taxonomic sampling coverage for rodent, shrew and hedgehogs. No single genera exceeds 50% of species within it sampled. Bar chart B) shows sampling biases by synanthropy states of speecies. NS refers to non-synanthropic species, OS are occasionally synanthropic species and TS are totally synanthropic. The majority of species fall in the unknown category. There is an increasing trend as you move from non-synanthropic through to totally synanthropic. Bar chart C) shows the number of unique viruses tested for by species. Most species are tested for 0-1 species, with a decay towards 10 viruses per species. There is a small secondary peak around 10 unique viruses suggesting some species are heavily investigated with the few species being tested for over 30 virus species.

Marginal effect plots of surveillance bias b) Night time lights, c) Remoteness and d) Host Richness. b) shows an increasing trend of sampling with more nighttime lights. c) shows a trend of decreasing sampling with increased remoteness (i.e., decreasing accessibility time/travel time from cities). d) shows no important effect of host species richness on sampling effort.

100

David Simons @davidsimons.bsky.social · 30/01/2025

Are any #Russian speakers are able to assist with this. Translation is from Google Translate does the original Russian state the species name or are they describing them similarly?

Machine translation from original russian text.

020