Bio Applied Variant Surveillance

Spike Protein Key Regions

Freyja Wastewater Pipeline

# 1. Call variants from wastewater BAM
freyja variants wastewater.bam --variants variants.tsv --depths depths.tsv

# 2. Deconvolve lineages (uses UShER barcodes)
freyja demix variants.tsv depths.tsv --output lineages.csv

# 3. Aggregate time-series
freyja aggregate --inputdir ./samples/ --output aggregated.tsv

# 4. Plot
freyja plot aggregated.tsv --output lineage_plot.pdf

Freyja model: linear mixture model solving for lineage fractions given observed variant frequencies and a barcode matrix of lineage-defining mutations.

Wastewater Epidemiology Facts

• Detects ~1 infected person per 100,000 catchment population
• Provides 3–7 day lead time before clinical case surge
• Population-level signal unaffected by clinical testing access gaps
• Signal = viral RNA copies/L (log-scale); smooth before correlation analysis

Python Patterns

import numpy as np
import pandas as pd
from scipy.stats import pearsonr
from scipy.ndimage import gaussian_filter1d

# Smooth wastewater signal before correlating with clinical cases
ww_smooth = gaussian_filter1d(viral_load, sigma=3)

# Lead-time correlation (WW leads cases by ~5 days/weeks)
r, p = pearsonr(ww_smooth[:-lag], cases[lag:])

# Lineage frequency normalization (stacked area plot)
freq_matrix = np.array(list(lineage_freqs.values()))
freq_norm = freq_matrix / freq_matrix.sum(axis=0) * 100

# Dominant lineage per time point
dominant_idx = np.argmax(freq_norm, axis=0)

Pitfalls

• Amplicon dropout: primer mismatches in divergent lineages → missing mutations → incorrect lineage calls; update primer schemes with each major variant wave
• Freyja barcode staleness: lineage barcodes must be updated regularly (freyja update); stale barcodes misclassify new sublineages as "Other"
• Wastewater normalization: viral load varies with flow rate and precipitation — normalize to pepper mild mottle virus (PMMoV) or crAssphage as fecal indicator
• Consensus vs quasispecies: Pangolin assigns one lineage to a consensus; co-infections or recombinants require manual inspection of per-site variant frequencies
• Coordinate systems: BED is 0-based; VCF/GFF are 1-based — off-by-one errors when mixing
• Multiple testing: apply Benjamini-Hochberg FDR when testing associations across many positions/lineages