@sjcrh/proteinpaint-server 2.177.0 → 2.178.0

package/routes/termdb.diffMeth.js

@@ -0,0 +1,180 @@
+import { diffMethPayload } from "#types/checkers";
+import { getData } from "../src/termdb.matrix.js";
+import { get_ds_tdb } from "../src/termdb.js";
+import { run_R } from "@sjcrh/proteinpaint-r";
+import { mayLog } from "#src/helpers.ts";
+import { formatElapsedTime } from "#shared";
+const api = {
+  endpoint: "termdb/diffMeth",
+  methods: {
+    get: {
+      ...diffMethPayload,
+      init
+    },
+    post: {
+      ...diffMethPayload,
+      init
+    }
+  }
+};
+function init({ genomes }) {
+  return async (req, res) => {
+    try {
+      const q = req.query;
+      const genome = genomes[q.genome];
+      if (!genome) throw new Error("invalid genome");
+      const [ds] = get_ds_tdb(genome, q);
+      let term_results = [];
+      if (q.tw) {
+        term_results = await getData({ filter: q.filter, filter0: q.filter0, terms: [q.tw] }, ds);
+        if (term_results.error) throw new Error(term_results.error);
+      }
+      let term_results2 = [];
+      if (q.tw2) {
+        term_results2 = await getData({ filter: q.filter, filter0: q.filter0, terms: [q.tw2] }, ds);
+        if (term_results2.error) throw new Error(term_results2.error);
+      }
+      const results = await run_diffMeth(req.query, ds, term_results, term_results2);
+      if (!results || !results.data) throw new Error("No data available");
+      res.send(results);
+    } catch (e) {
+      res.send({ status: "error", error: e.message || e });
+      if (e instanceof Error && e.stack) console.log(e);
+    }
+  };
+}
+async function run_diffMeth(param, ds, term_results, term_results2) {
+  if (param.samplelst?.groups?.length != 2) throw new Error(".samplelst.groups.length!=2");
+  if (param.samplelst.groups[0].values?.length < 1) throw new Error("samplelst.groups[0].values.length<1");
+  if (param.samplelst.groups[1].values?.length < 1) throw new Error("samplelst.groups[1].values.length<1");
+  const q = ds.queries.dnaMethylation?.promoter;
+  if (!q) throw new Error("ds.queries.dnaMethylation.promoter is not configured");
+  if (!q.file) throw new Error("ds.queries.dnaMethylation.promoter.file is missing");
+  const group1names = [];
+  const conf1_group1 = [];
+  const conf2_group1 = [];
+  for (const s of param.samplelst.groups[0].values) {
+    if (!Number.isInteger(s.sampleId)) continue;
+    const n = ds.cohort.termdb.q.id2sampleName(s.sampleId);
+    if (!n) continue;
+    if (!q.allSampleSet.has(n)) continue;
+    if (param.tw && param.tw2) {
+      if (term_results.samples[s.sampleId] && term_results2.samples[s.sampleId]) {
+        conf1_group1.push(
+          param.tw.q.mode == "continuous" ? term_results.samples[s.sampleId][param.tw.$id]["value"] : term_results.samples[s.sampleId][param.tw.$id]["key"]
+        );
+        conf2_group1.push(
+          param.tw2.q.mode == "continuous" ? term_results2.samples[s.sampleId][param.tw2.$id]["value"] : term_results2.samples[s.sampleId][param.tw2.$id]["key"]
+        );
+        group1names.push(n);
+      }
+    } else if (param.tw && !param.tw2) {
+      if (term_results.samples[s.sampleId]) {
+        conf1_group1.push(
+          param.tw.q.mode == "continuous" ? term_results.samples[s.sampleId][param.tw.$id]["value"] : term_results.samples[s.sampleId][param.tw.$id]["key"]
+        );
+        group1names.push(n);
+      }
+    } else if (!param.tw && param.tw2) {
+      if (term_results2.samples[s.sampleId]) {
+        conf2_group1.push(
+          param.tw2.q.mode == "continuous" ? term_results2.samples[s.sampleId][param.tw2.$id]["value"] : term_results2.samples[s.sampleId][param.tw2.$id]["key"]
+        );
+        group1names.push(n);
+      }
+    } else {
+      group1names.push(n);
+    }
+  }
+  const group2names = [];
+  const conf1_group2 = [];
+  const conf2_group2 = [];
+  for (const s of param.samplelst.groups[1].values) {
+    if (!Number.isInteger(s.sampleId)) continue;
+    const n = ds.cohort.termdb.q.id2sampleName(s.sampleId);
+    if (!n) continue;
+    if (!q.allSampleSet.has(n)) continue;
+    if (param.tw && param.tw2) {
+      if (term_results.samples[s.sampleId] && term_results2.samples[s.sampleId]) {
+        conf1_group2.push(
+          param.tw.q.mode == "continuous" ? term_results.samples[s.sampleId][param.tw.$id]["value"] : term_results.samples[s.sampleId][param.tw.$id]["key"]
+        );
+        conf2_group2.push(
+          param.tw2.q.mode == "continuous" ? term_results2.samples[s.sampleId][param.tw2.$id]["value"] : term_results2.samples[s.sampleId][param.tw2.$id]["key"]
+        );
+        group2names.push(n);
+      }
+    } else if (param.tw && !param.tw2) {
+      if (term_results.samples[s.sampleId]) {
+        conf1_group2.push(
+          param.tw.q.mode == "continuous" ? term_results.samples[s.sampleId][param.tw.$id]["value"] : term_results.samples[s.sampleId][param.tw.$id]["key"]
+        );
+        group2names.push(n);
+      }
+    } else if (!param.tw && param.tw2) {
+      if (term_results2.samples[s.sampleId]) {
+        conf2_group2.push(
+          param.tw2.q.mode == "continuous" ? term_results2.samples[s.sampleId][param.tw2.$id]["value"] : term_results2.samples[s.sampleId][param.tw2.$id]["key"]
+        );
+        group2names.push(n);
+      }
+    } else {
+      group2names.push(n);
+    }
+  }
+  const sample_size1 = group1names.length;
+  const sample_size2 = group2names.length;
+  const alerts = validateGroups(sample_size1, sample_size2, group1names, group2names);
+  if (param.preAnalysis) {
+    const group1Name = param.samplelst.groups[0].name;
+    const group2Name = param.samplelst.groups[1].name;
+    return {
+      data: {
+        [group1Name]: sample_size1,
+        [group2Name]: sample_size2,
+        ...alerts.length ? { alert: alerts.join(" | ") } : {}
+      }
+    };
+  }
+  if (alerts.length) throw new Error(alerts.join(" | "));
+  const diffMethInput = {
+    case: group2names.join(","),
+    control: group1names.join(","),
+    input_file: q.file,
+    min_samples_per_group: param.min_samples_per_group
+  };
+  if (param.tw) {
+    diffMethInput.conf1 = [...conf1_group2, ...conf1_group1];
+    diffMethInput.conf1_mode = param.tw.q.mode;
+    if (new Set(diffMethInput.conf1).size === 1) {
+      throw new Error("Confounding variable 1 has only one value");
+    }
+  }
+  if (param.tw2) {
+    diffMethInput.conf2 = [...conf2_group2, ...conf2_group1];
+    diffMethInput.conf2_mode = param.tw2.q.mode;
+    if (new Set(diffMethInput.conf2).size === 1) {
+      throw new Error("Confounding variable 2 has only one value");
+    }
+  }
+  const time1 = Date.now();
+  const result = JSON.parse(await run_R("diffMeth.R", JSON.stringify(diffMethInput)));
+  mayLog("Time taken to run diffMeth:", formatElapsedTime(Date.now() - time1));
+  const output = {
+    data: result.promoter_data,
+    sample_size1,
+    sample_size2
+  };
+  return output;
+}
+function validateGroups(sample_size1, sample_size2, group1names, group2names) {
+  const alerts = [];
+  if (sample_size1 < 1) alerts.push("sample size of group1 < 1");
+  if (sample_size2 < 1) alerts.push("sample size of group2 < 1");
+  const commonnames = group1names.filter((x) => group2names.includes(x));
+  if (commonnames.length) alerts.push(`Common elements found between both groups: ${commonnames.join(", ")}`);
+  return alerts;
+}
+export {
+  api
+};

package/routes/termdb.dmr.js

@@ -0,0 +1,48 @@
+import { TermdbDmrPayload } from "#types/checkers";
+import { run_R } from "@sjcrh/proteinpaint-r";
+import { invalidcoord } from "#shared/common.js";
+const api = {
+  endpoint: "termdb/dmr",
+  methods: {
+    get: {
+      ...TermdbDmrPayload,
+      init
+    },
+    post: {
+      ...TermdbDmrPayload,
+      init
+    }
+  }
+};
+function init({ genomes }) {
+  return async (req, res) => {
+    try {
+      const q = req.query;
+      const genome = genomes[q.genome];
+      if (!genome) throw new Error("invalid genome");
+      const ds = genome.datasets?.[q.dslabel];
+      if (!ds) throw new Error("invalid ds");
+      if (!ds.queries?.dnaMethylation) throw new Error("not supported");
+      if (!Array.isArray(q.group1) || q.group1.length == 0) throw new Error("group1 not non empty array");
+      if (!Array.isArray(q.group2) || q.group2.length == 0) throw new Error("group2 not non empty array");
+      if (invalidcoord(genome, q.chr, q.start, q.stop)) throw new Error("invalid chr/start/stop");
+      const arg = {
+        group1: q.group1,
+        group2: q.group2,
+        file: ds.queries.dnaMethylation.file,
+        // todo change file to mValueFile
+        chr: q.chr,
+        start: q.start,
+        stop: q.stop
+      };
+      const result = JSON.parse(await run_R("dmr.R", JSON.stringify(arg)));
+      if (result.error) throw new Error(result.error);
+      res.send(result);
+    } catch (e) {
+      res.send({ error: e.message || e });
+    }
+  };
+}
+export {
+  api
+};

package/routes/termdb.sampleScatter.js

@@ -119,9 +119,15 @@ function init({ genomes }) {
 async function getSingleCellScatter(req, res, ds) {
   const q = req.query;
   const { name, sample } = q.singleCellPlot;
-  const tw = q.colorTW;
   try {
-    const data = await ds.queries.singleCell.data.get({ plots: [name], sample });
+    const tw = q.colorTW;
+    const arg = { plots: [name], sample };
+    if (tw) {
+      if (tw.term.type == "singleCellGeneExpression") arg.gene = tw.term.gene;
+      else if (tw.term.type == "singleCellCellType") arg.colorBy = tw.term.name;
+      else throw new Error("unsupported tw");
+    }
+    const data = await ds.queries.singleCell.data.get(arg);
     const plot = data.plots[0];
     const cells = [...plot.expCells, ...plot.noExpCells];
     const samples = cells.map((cell) => {
@@ -464,6 +470,10 @@ async function loadFile(p, ds) {
   );
 }
 async function mayInitiateScatterplots(ds) {
+  if (ds.scatterplots) {
+    ds.cohort.scatterplots = ds.scatterplots;
+    delete ds.scatterplots;
+  }
   if (!ds.cohort.scatterplots) return;
   if (typeof ds.cohort.scatterplots.get == "function") {
   }

package/routes/termdb.singlecellSamples.js

@@ -9,6 +9,7 @@ import { termdbSingleCellSamplesPayload } from "#types/checkers";
 import { validate_query_singleCell_DEgenes } from "./termdb.singlecellDEgenes.ts";
 import { gdc_validate_query_singleCell_data } from "#src/mds3.gdc.js";
 import ky from "ky";
+import { TermTypes } from "#shared/terms.js";
 const api = {
   endpoint: "termdb/singlecellSamples",
   methods: {
@@ -59,6 +60,7 @@ async function validate_query_singleCell(ds, genome) {
   } else {
     throw new Error("unknown singleCell.data.src");
   }
+  colorColumn2terms(ds.queries.singleCell.data.plots, ds);
   if (q.geneExpression) {
     if (typeof q.geneExpression != "object") throw new Error("singleCell.geneExpression not object");
     if (q.geneExpression.src == "native") {
@@ -98,6 +100,7 @@ async function validateSamplesNative(S, D, ds) {
       if (sid == void 0) throw new Error(`singlecell.sample: unknown sample name ${sampleName}`);
       samples.set(sid, { sample: sampleName });
     }
+    if (!plot.colorColumns || plot.colorColumns.length == 0) continue;
   }
   if (S.sampleColumns) {
     for (const { termid } of S.sampleColumns) {
@@ -126,11 +129,16 @@ function validateDataNative(D, ds) {
     const plots = [];
     let geneExpMap;
     if (ds.queries.singleCell.geneExpression && q.gene) {
-      geneExpMap = await ds.queries.singleCell.geneExpression.get({ sample: q.sample, gene: q.gene });
+      const sample = q.sample || q.singleCellPlot.sample;
+      geneExpMap = await ds.queries.singleCell.geneExpression.get({ sample, gene: q.gene });
     }
     for (const plot of D.plots) {
       if (!q.plots.includes(plot.name)) continue;
-      const tsvfile = path.join(serverconfig.tpmasterdir, plot.folder, (q.sample.eID || q.sample.sID) + plot.fileSuffix);
+      const tsvfile = path.join(
+        serverconfig.tpmasterdir,
+        plot.folder,
+        (q.sample?.eID || q.sample?.sID) + plot.fileSuffix
+      );
       if (!file2Lines[tsvfile]) {
         await file_is_readable(tsvfile);
         const text = await read_file(tsvfile);
@@ -146,7 +154,7 @@ function validateDataNative(D, ds) {
         }
         file2Lines[tsvfile] = lines2;
       }
-      const colorColumn = plot.colorColumns.find((c) => c.name == q.colorBy?.[plot.name]) || plot.colorColumns[0];
+      const colorColumn = plot.colorColumns.find((c) => c.name == q.colorBy) || plot.colorColumns[0];
       const expCells = [];
       const noExpCells = [];
       for (const l of file2Lines[tsvfile]) {
@@ -182,7 +190,7 @@ function validateDataNative(D, ds) {
 function validateGeneExpressionNative(G) {
   G.sample2gene2expressionBins = {};
   G.get = async (q) => {
-    const h5file = path.join(serverconfig.tpmasterdir, G.folder, (q.sample.eID || q.sample.sID) + ".h5");
+    const h5file = path.join(serverconfig.tpmasterdir, G.folder, (q.sample?.eID || q.sample?.sID) + ".h5");
     await file_is_readable(h5file);
     const query_gene = q.gene;
     if (!query_gene) {
@@ -233,6 +241,34 @@ function gdc_validateGeneExpression(G, ds, genome) {
     }
   };
 }
+function colorColumn2terms(plots, ds) {
+  const termSet = /* @__PURE__ */ new Set();
+  for (const plot of plots) {
+    const tmpTerms = plot.colorColumns.map((c) => {
+      const baseValues = c.colorMap ? Object.keys(c.colorMap) : [];
+      return {
+        name: c.name,
+        isleaf: true,
+        /** TODO: possible term may apply to multiple plots.
+         * May need to change to plots: [] */
+        plot: plot.name,
+        type: TermTypes.SINGLECELL_CELLTYPE,
+        groupsetting: {},
+        values: baseValues.reduce((acc, v) => {
+          const alias = c?.aliases?.[v];
+          acc[v] = {
+            key: v,
+            label: alias || v,
+            color: c.colorMap?.[v] || "#000000"
+          };
+          return acc;
+        }, {})
+      };
+    });
+    tmpTerms.forEach((term) => termSet.add(term));
+  }
+  ds.queries.singleCell.terms = [...termSet];
+}
 export {
   api,
   validate_query_singleCell