Variant Comparison — Relative Growth
cov-spectrum style pairwise comparison: weekly share of B within the A+B pool, log-odds regressed over ISO weeks. A positive advantage means B is outgrowing A.
A/B inputs accept three syntaxes: a plain lineage (SW.2.1, SW.2.1* = with sublineages); a cov-spectrum style expression with + (AND) and ! (NOT) — e.g. PQ.16.1.1*+S:H445Q, PQ.16.1.1*+!S:H445Q, !S:H445Q alone; or a full regex over each sequence's lineage,mutations string — e.g. ^PQ\.16\.1\.1(\.|,)(?=.*S:H445Q). Mutation tokens are uppercase without AA:/NUC: prefixes (S:V445Q, T3090C) and relative to the Wuhan reference — a des.json label like SV.2.4's S:H445Q (parent-relative) is stored here as S:V445Q.
ST.1* grows +45.1%/wk faster than SW.2.1* (95% CI +39.6% to +50.8%)
Share of B within the pair
Weekly pair share
| ISO week | A: SW.2.1* | B: ST.1* | A+B | B share of pair | In fit |
|---|---|---|---|---|---|
| 2026-W31 | 82 | 5 | 87 | 5.8% | ✓ |
| 2026-W32 | 110 | 17 | 127 | 13.4% | ✓ |
| 2026-W33 | 110 | 20 | 130 | 15.4% | ✓ |
| 2026-W34 | 141 | 58 | 199 | 29.1% | ✓ |
| 2026-W35 | 144 | 65 | 209 | 31.1% | ✓ |
| 2026-W36 | 141 | 87 | 228 | 38.2% | ✓ |
| 2026-W37 | 132 | 118 | 250 | 47.2% | ✓ |
| 2026-W38 | 100 | 150 | 250 | 60.0% | ✓ |
| 2026-W39 | 105 | 177 | 282 | 62.8% | ✓ |
| 2026-W40 | 90 | 164 | 254 | 64.6% | ✓ |
| 2026-W41 | 19 | 101 | 120 | 84.2% | ✓ |
| 2026-W42 | 3 | 18 | 21 | 85.7% | ✓ |
| 2026-W43 | - | 3 | 3 | 100.0% | - |
Pairwise advantage matrix
Each cell: weekly growth advantage of the row variant over the column variant (B vs A baseline). Green = row outgrows column. Click a cell to open that pair.
| row vs column | XFG | NB.1.8.1 | XFG.6 | JN.1 | PQ.16.1.1 | XFG.23.1.3 | SW.2.1 | RF.5 |
|---|---|---|---|---|---|---|---|---|
| XFG | — | +15% | -2% | -3% | +14% | +1% | +12% | +14% |
| NB.1.8.1 | -13% | — | -15% | -16% | -1% | -12% | -3% | -1% |
| XFG.6 | +2% | +18% | — | -1% | +17% | +4% | +15% | +16% |
| JN.1 | +4% | +19% | +1% | — | +18% | +5% | +16% | +18% |
| PQ.16.1.1 | -12% | +1% | -14% | -15% | — | -11% | -4% | -6% |
| XFG.23.1.3 | -1% | +14% | -4% | -5% | +13% | — | +11% | +8% |
| SW.2.1 | -11% | +3% | -13% | -14% | +4% | -10% | — | -2% |
| RF.5 | -12% | +1% | -14% | -15% | +6% | -7% | +2% | — |
Method: for each ISO week, share = B/(A+B); y = log((B+0.5)/(A+0.5)) regressed on week index; advantage = exp(slope)−1 per week. Weeks with A+B below the threshold are excluded from the fit (shown hollow in the chart). This is a relative measure — both variants can be declining in absolute terms while B still shows an advantage over A.