PneumoTRACE

Enter the access password to continue.

Incorrect password. Please try again.

Serotype patterns in carriage

Carriage data

This page shows the serotype distribution by individual characteristics, including demographics and vaccine history

Serotype frequency by age group

This shows the proportion of carriage episodes in each age group with each serotype. Only samples with a valid serotype call are included in these totals.

The most important serotypes in children <5y, the most commonly detected serotypes were 23B, 35B,23A,31,35F, 19F, 21, 6C, 7C.

This list largely agrees with recently presented conference findings from Rochester (top serotypes, in order were 35B, 23B, 23A, 19F, 35F, 11A, 15C, 6C) and Boston (35B, 23A, 23B, 35F,21, 15B, 15A, 11A, 22F). Note that we detect a lot of serotype 43, but we are confident that serotype 43 is an artifact of the sequencing assay and are working on a solution to clean it up.

Important note: these serotype frequencies should be considered provisional. We are still working on improving the sequencing assay, and some serotypes, particularly within serogroups, can be difficult to resolve. We will be working to improve on the sequencing assays to improve both specificity and sensitivity.

Serotype	N	Proportion of serotyped isolates
Serotype distribution in children <5y
23B	17	16.8%
35B	14	13.9%
23A	7	6.9%
(43*)	7	6.9%
31	6	5.9%
35F	6	5.9%
19F	5	5.0%
21	5	5.0%
6C	4	4.0%
7C	4	4.0%
11D	3	3.0%
15BC	3	3.0%
45	3	3.0%
11A	2	2.0%
11E	2	2.0%
35D	2	2.0%
5	2	2.0%
15A	1	1.0%
16A	1	1.0%
22F	1	1.0%
23F	1	1.0%
33H	1	1.0%
34	1	1.0%
7D	1	1.0%
9A	1	1.0%
9V	1	1.0%
*Serotype 43 is likely a lab artifact

alluvialData = transpose(alluvialRaw)
stCats = stOrder
ageCats = ageOrder

// Stable colour per serotype, "Other" always gray; concatenated categorical
// schemes so a double-digit number of serotypes still get distinct colours.
alluPalette = (() => {
  const base = [].concat(d3.schemeTableau10, d3.schemeSet3, d3.schemePaired)
  const named = stCats.slice(0, -1)
  const m = new Map(named.map((s, i) => [s, base[i % base.length]]))
  m.set("Other", "#c9ccd1")
  return m
})()

// Stacked band [y0, y1] (0-1 share) per age group x serotype. Stacks from the
// top down in stCats order (matching ggalluvial's default in the static
// chart), so the same serotype occupies a traceable band across ages.
//
// A serotype absent from an age group is floored to a thin sliver (rather
// than 0) so its flow still traces a faint continuous line instead of a gap;
// the floored shares are then renormalized so each stack still sums to 100%.
alluBands = {
  const MIN_VISUAL_PCT = 0.003
  const byAge = new Map()
  for (const a of ageCats) {
    const rows = stCats.map(s => {
      const row = alluvialData.find(d => d.agec === a && d.st_grp === s)
      const pct = row ? row.pct : 0
      const n = row ? row.n : 0
      return { s, pct, n, plotPct: Math.max(pct, MIN_VISUAL_PCT) }
    })
    const total = d3.sum(rows, d => d.plotPct)
    const bands = new Map()
    let yTop = 1
    for (const r of rows) {
      const share = r.plotPct / total
      const y1 = yTop, y0 = yTop - share
      bands.set(r.s, { y0, y1, pct: r.pct, n: r.n })
      yTop = y0
    }
    byAge.set(a, bands)
  }
  return byAge
}

alluvialChart = {
  const W = Math.min(width, 900), H = 560
  const marginLeft = 50, marginRight = 12, marginTop = 26, marginBottom = 34
  const innerW = W - marginLeft - marginRight, innerH = H - marginTop - marginBottom
  const n = ageCats.length
  const bandGap = innerW / n
  const bandW = Math.min(70, bandGap * 0.42)
  const xCenter = ageCats.map((a, i) => marginLeft + bandGap * (i + 0.5))
  const yOf = pct => marginTop + (1 - pct) * innerH

  const svg = d3.create("svg")
    .attr("viewBox", [0, 0, W, H]).attr("width", W).attr("height", H)
    .style("font", "11px sans-serif");

  // gridlines + percent axis
  const gridG = svg.append("g");
  [0, 0.25, 0.5, 0.75, 1].forEach(p => {
    gridG.append("line")
      .attr("x1", marginLeft).attr("x2", W - marginRight)
      .attr("y1", yOf(p)).attr("y2", yOf(p)).attr("stroke", "#e3e3e3");
    gridG.append("text")
      .attr("x", marginLeft - 8).attr("y", yOf(p)).attr("dy", "0.32em")
      .attr("text-anchor", "end").attr("fill", "#555")
      .text(d3.format(".0%")(p));
  });
  svg.append("text")
    .attr("transform", "rotate(-90)")
    .attr("x", -(marginTop + innerH / 2)).attr("y", 14)
    .attr("text-anchor", "middle").attr("font-weight", 700).attr("font-size", 12)
    .attr("fill", "#132237")
    .text("Percent of typed samples");

  // age-group axis labels
  ageCats.forEach((a, i) => {
    svg.append("text")
      .attr("x", xCenter[i]).attr("y", H - 8)
      .attr("text-anchor", "middle").attr("font-weight", 600).attr("fill", "#132237")
      .text(a);
  });

  // floating tooltip, positioned relative to the wrapper div
  const tip = html`<div style="position:absolute; pointer-events:none; opacity:0;
    background:rgba(19,34,55,0.95); color:#eff0f1; font:12px sans-serif;
    padding:5px 8px; border-radius:4px; transition:opacity 0.1s; z-index:10;
    white-space:nowrap;"></div>`;
  const wrapper = html`<div style="position:relative;">${tip}</div>`;

  function showTip(event, text) {
    tip.innerHTML = text;
    tip.style.opacity = 1;
    moveTip(event);
  }
  function moveTip(event) {
    const [mx, my] = d3.pointer(event, wrapper);
    tip.style.left = `${mx + 12}px`;
    tip.style.top = `${my - 10}px`;
  }
  function hideTip() { tip.style.opacity = 0; }

  // flow ribbons between each pair of adjacent age groups
  const flowG = svg.append("g");
  for (let i = 0; i < n - 1; i++) {
    const a0 = ageCats[i], a1 = ageCats[i + 1];
    const x0 = xCenter[i] + bandW / 2, x1 = xCenter[i + 1] - bandW / 2, xm = (x0 + x1) / 2;
    for (const s of stCats) {
      const b0 = alluBands.get(a0).get(s), b1 = alluBands.get(a1).get(s);
      const path = d3.path();
      path.moveTo(x0, yOf(b0.y0));
      path.bezierCurveTo(xm, yOf(b0.y0), xm, yOf(b1.y0), x1, yOf(b1.y0));
      path.lineTo(x1, yOf(b1.y1));
      path.bezierCurveTo(xm, yOf(b1.y1), xm, yOf(b0.y1), x0, yOf(b0.y1));
      path.closePath();
      flowG.append("path")
        .attr("d", path.toString())
        .attr("fill", alluPalette.get(s)).attr("fill-opacity", 0.55)
        .on("mouseenter", function (event) {
          d3.select(this).attr("fill-opacity", 0.85);
          showTip(event, `<b>${s}</b><br>${a0}: ${d3.format(".1%")(b0.pct)}<br>${a1}: ${d3.format(".1%")(b1.pct)}`);
        })
        .on("mousemove", moveTip)
        .on("mouseleave", function () {
          d3.select(this).attr("fill-opacity", 0.55);
          hideTip();
        });
    }
  }

  // stratum blocks, each directly labeled with its serotype
  const stratG = svg.append("g");
  ageCats.forEach((a, i) => {
    const x = xCenter[i];
    for (const s of stCats) {
      const b = alluBands.get(a).get(s);
      stratG.append("rect")
        .attr("x", x - bandW / 2).attr("y", yOf(b.y1))
        .attr("width", bandW).attr("height", Math.max(0.5, yOf(b.y0) - yOf(b.y1)))
        .attr("fill", alluPalette.get(s))
        .on("mouseenter", function (event) {
          d3.select(this).attr("stroke", "#132237").attr("stroke-width", 1.5);
          showTip(event, `<b>${s}</b><br>${a}: ${d3.format(".1%")(b.pct)} (n=${b.n})`);
        })
        .on("mousemove", moveTip)
        .on("mouseleave", function () {
          d3.select(this).attr("stroke", "none");
          hideTip();
        });
      if ((b.y1 - b.y0) * innerH > 12) {
        stratG.append("text")
          .attr("x", x).attr("y", (yOf(b.y0) + yOf(b.y1)) / 2).attr("dy", "0.32em")
          .attr("text-anchor", "middle").attr("font-weight", 700).attr("font-size", 10.5)
          .attr("fill", s === "Other" ? "#333" : "#132237")
          .attr("pointer-events", "none")
          .text(s);
      }
    }
  });

  wrapper.prepend(svg.node());
  return wrapper;
}

Observed vs. expected carriage frequency

This plot compares the observed frequency of serotypes in carriage samples to the frequency we would expect based on 2024 IPD surveillance data and published invasiveness estimates. Serotypes above the diagonal are more common in carriage than expected, while those below are less common. This comparison helps identify (1) if there are adjustments that need to be made to the sequencing assay (e.g., it is unlikely that the serotype 43 carriage is a real signal) or (2) if there are notable changes. These comparisons will become more valuable over time as we are able to distinguish assay issues from true signal.

Serotype frequency by presence of a virus

Serotype distribution among pneumococcal-positive samples, split by whether a virus was also detected (combined across age groups). Positive samples that did not receive a confident serotype call are shown as their own row rather than excluded.

Serotype	Virus Negative		Virus Positive
Serotype distribution among pneumococcal-positive samples, by viral status
Serotype	N	% of positives	N	% of positives
43	11	1.1%	12	2.0%
23B	9	0.9%	13	2.1%
35B	9	0.9%	11	1.8%
23A	7	0.7%	8	1.3%
15BC	2	0.2%	5	0.8%
31	3	0.3%	4	0.7%
35F	1	0.1%	6	1.0%
7C	3	0.3%	4	0.7%
11A	2	0.2%	4	0.7%
19F	1	0.1%	5	0.8%
Other serotype	27	2.8%	32	5.3%
Uncolonized/no call	895	92.3%	502	82.8%

Serotype frequency by age and immunization history

Coming soon!

Which vaccines did the individual received?

Coming soon!

Serotype frequency by age and population density of the ZIP of residence

Coming soon!

Serotype frequency by age and race/ethnicity

Coming soon!

Serotype frequency by age and reason for the visit

Coming soon!