The disease forecast is getting granular. The pathogens are moving too.

Heat turns the resistance dial

Start with a preprint that makes warming concretely biological. In E. coli, the selection strength for kanamycin resistance rises with temperature, driven by faster growth and greater susceptibility — heat makes resistance pay off more. Plasmid transfer tells a different story: the R1 plasmid's transfer rate has a narrower thermal tolerance and peaks at 35C, well below the 40C bacterial growth peak. And in mixed populations, the spread of resistance is driven mostly by selection, not by transfer. Put plainly: in a warmer world, resistance spreads because bacteria grow faster and die more easily when susceptible, not because they trade plasmids more eagerly. The policy lever is the temperature-selection gradient, not the conjugation rate. m:…3394

Risk, down to the county

The temperature detail matters for how the problem gets modeled. Transfer and selection peak at different temperatures, so a single warming-helps-resistance coefficient would mislead. And the mixed-population result — selection, not transfer, driving spread — redirects attention away from plasmid-blocking interventions and toward whatever lowers selection pressure as conditions warm. Heat is not just making more bacteria; it is making resistance a better bet. m:…3394

Forecasting is also going finer-grained. A preprint mapping tick-borne disease risk across Illinois counties combines vector abundance, pathogen prevalence, and phenology into ecological risk scores — and those scores track actual human case rates for three tick species: A. americanum, D. variabilis, and I. scapularis. The county rankings hold steady under moderate changes to how the inputs are weighted, which is the kind of robustness a public health department needs before acting on a map. I. scapularis, the Lyme vector, shows a broader seasonal risk profile than the other two, whose risk concentrates in the warm season. A county health official can read this as a calendar, not just a map. m:…9550

Dengue planning is getting the same granular treatment. A preprint reconstructing dengue exposure dynamics finds that spatial heterogeneity in accumulated immunity is closely associated with projected disease burden and with where vaccines should go. Targeting children produces the greatest reductions in disability-adjusted life years; targeting the elderly produces smaller DALY gains but larger reductions in mortality. And accelerating vaccine supply markedly increases the program's public health impact. The allocation question is no longer just who, but who and where — and the answer depends on the immune history already written into each district. m:…2486

The moving target

The common thread across the maps is resolution. Risk scores that survive re-weighting, allocations that depend on local immune history, seasonal profiles that differ by species, selection gradients that shift with temperature — forecasting is moving from national curves to county calendars. And the models are honest about where the leverage sits: supply speed for dengue, weighting stability for ticks, selection pressure for resistance. Granularity is not precision theater; it is where the decisions live. m:…9550 m:…2486 m:…3394

Here is the catch, and it arrives in a preprint on SARS-CoV-2. The spike receptor-binding domain has evolved decreased inherent antigenicity in humans: a smaller fraction of naive human B cells bind the spike protein than before, with mutations E484A and F486P doing the work. The effect is human-specific — it does not show up in mouse models, which is a quiet warning about how much we lean on them. The virus is getting harder for a first-time immune system to notice. All the careful mapping in the world still has to contend with a target that keeps redesigning its surface. The forecast gets granular; the pathogen gets slippery. Both trends are true at once, and any serious infectious-disease planning has to hold them together. m:…8399