Status
Preliminary — data extraction and coordinate retrieval complete (NB01). NB02 spatial analysis and NB03 figures pending.
Key Findings
- 260,652 environmental MAGs extracted from MGnify; 30,497 after filtering host-associated biomes.
- 22,356 MAGs with usable geospatial coordinates (73.3% coordinate coverage of environmental MAG set).
- ENA batch API retrieval: 24,511 sample records; 16,964 (69.2%) have valid lat/lon pairs — a 30.8% spatial data gap in public metagenomic archives.
- Global map code drafted (geopandas + matplotlib) showing distribution of metal-resistant (n_metal_types > 0) vs. susceptible MAGs coloured by metal diversity.
- Known issue: NameError on plt import in map cell (missing
import matplotlib.pyplotin that cell); fix required before NB02.
Interpretation
Roughly one third of publicly archived metagenomic samples lack usable geospatial coordinates, constituting a systematic gap in the ability to map microbial metal resistance at global scale. Among the 22,356 MAGs with coordinates, metal-resistant genomes are not uniformly distributed, but hotspot identification and sampling-effort correction are required before regional patterns can be interpreted (NB02).
The 30.8% ENA coordinate gap is a data quality finding independent of the biological question: it implies that any global map of metal resistance distribution based on public archives is subject to substantial geographic blind spots. Which biomes are most underrepresented is a key open question.
NB02 Spatial Analysis Results
Global metal resistance prevalence: 2.8% of 22,356 MAGs with coordinates carry ≥1 metal resistance type — substantially lower than implied by reporting on the 30,497 filtered MAG set. Metal resistance is rare in the global environmental MAG pool.
5° grid hotspot analysis (289 cells with ≥5 MAGs; Fisher’s exact, BH FDR):
- 11 significant hotspots (OR>2, q<0.05), 3 coldspots (OR<0.5, q<0.05).
- Top hotspot: Atacama/Andean region, Chile (lat=-25°, lon=-70°): 21.8% prevalence, OR=9.83, q=7.6e-12, n=101 MAGs.
- Eastern and central USA clusters (lat=40°, lon=-80° and -90°): OR=7.9 and 6.3.
- East/Southeast Asia cluster (lat=25-30°, lon=105-120°): OR=4.4–5.9.
Figure:
figures/fig_nb02_global_hotspot_map.png
Biome-stratified prevalence (Fisher’s exact vs. global baseline, BH FDR):
| Biome | n MAGs | Prevalence | OR | q |
|---|---|---|---|---|
| Soil | 7,939 | 5.8% | 5.05 | 1.8e-82 |
| Rhizosphere | 422 | 1.9% | 0.66 | 0.30 (NS) |
| Marine | 13,995 | 1.2% | 0.20 | 9.2e-80 |
Soil is significantly enriched for metal resistance (5.8%, OR=5.05); Marine is
significantly depleted (1.2%, OR=0.20). Figure: figures/fig_nb02_biome_prevalence.png
ENA coordinate gap: 0 of 532 MAG grid cells lack ENA coordinate coverage — the ENA coordinate file already covers all areas represented in the MAG coordinate dataset. The 30.8% gap (16,964/24,511 valid pairs) is a per-sample gap, not a geographic gap; it reflects samples without coordinates, not grid cells with no samples at all.
Key corrections to REPORT:
- The “22,356 MAGs with coordinates” dataset shows 2.8% global metal resistance prevalence, not a higher figure. This is much lower than the 21.8% T4SS prevalence, confirming that metal resistance genes (AMRFinderPlus) and HGT machinery (T4SS) are distinct features not uniformly co-distributed.
- Expedition-level clustering check pending for the Atacama and USA clusters.
Pending Analyses
- Expedition-level clustering: test if Atacama (OR=9.8) and USA clusters are single-study artefacts (check distinct sample_accession prefixes per hotspot)
- Sampling effort correction: confirm hotspots persist after normalising by log(n_MAGs)
- Fix matplotlib import in NB01 map cell for production figure