31:T779,table (by_gene[0], { limit: 5000, enableCSVDownload: true, sort: true, enableFilter: false, columns: [ { key: 'gene', name: 'Gene/ORF' }, { key: 'sequences', name: 'Seqs' }, { key: 'total', name: 'Sites' }, { key: 'variable', name: "Sites w/variants", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'variable-aa', name: "AA variants", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'variable-2x', name: "Shared variants", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'variable-2x-aa', name: "Shared AA variants", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'selected-positive', name: "+ selection", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'selected-negative', name: "- selection", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'in_epitope', name: "HLA epitope", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'in_epitope_positive', name: "HLA epitope, +", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] }, { key: 'in_epitope_negative', name: "HLA epitope, -", render: (val)=>per_kp ? d3.format ("1g")(val[1]) : val[0] } ] });32:Tcf1,genome_view = vegalite({ "$schema": "https://vega.github.io/schema/vega-lite/v5.json", "config": { "view": { "stroke": null }, }, "data": { "values": _.filter (by_gene_plot, (r)=>r.gene == gene) }, "vconcat": [{ "height": 120, "width": 750, "mark": { type: "tick", filled : true, tooltip: true }, "encoding": { "y": { "field": "kind", "type": "nominal", "sort": null, "scale" : { "domain" : ['Nucleotide', 'Nucleotide 2x', 'AA SNP', 'Shared AA SNP','+ selection in SARS-CoV-2', '+ selection in beta-CoV', '- selection in SARS-CoV-2', '- selection in beta-CoV','Unexpected codons'] }, "axis": { "domain": false, "offset": 10, "labelFontSize": "14", "ticks": false, "grid": true, "title": null } }, "size" : { "value" :10, }, "color" : { "field" : "type", "scale" : { "domain" : ["Conserved", "Variable", "Adaptive"], }, "legend" : {"title" : "Site kind", "orient" : "bottom", "title" : "", "labelFontSize": "14",} }, "x": { "field": "coord", "type": "quantitative", "scale": { /*"domain": { "selection": "brush" }*/ }, "axis": { "title": "Codon site in " + gene, "grid": false, "orient" : "top", "titleFontSize": "14", "labelFontSize": "12", } }, } }, /*{ "height": 40, "width": 750, "layer": [{ "selection": { "brush": { "type": "interval", "encodings": ["x"], "mark": { "fill": "#fdbb84", "fillOpacity": 0.5, "stroke": "#e34a33" } } }, "transform": [{ "calculate": "3", "as": "y" }, { "calculate": "12", "as": "y2" }], "mark": { type: "rule" }, "encoding": { "tooltip": [{"field": "position", "type": "quantitative"}, {"field": "gene", "type": "nominal"}, {"field": "site_in_gene", "type": "quantitative"}, {"field": "score", "type": "quantitative"}, {"field": "type", "type": "nominal"}, {"field": "peptide", "type": "nominal"},], "x": { "field": "position", "type": "quantitative", "axis": null, }, "y": { "field": "y", "type": "quantitative", "axis": null, }, "y2": { "field": "y2", "type": "quantitative", "axis": null, }, "color": { "field": "score", "type": "nominal", "legend": { "title": "Site score", "orient": "bottom" }, "scale": { "scheme": "lightgreyred" } } } } ] },*/ ] })33:Td30,counts_scatter = vegalite({ "title": null, data: {values: by_gene[1]}, "vconcat" : _.map (["variable","variable-2x","variable-aa","variable-2x-aa","selected-positive","selected-negative"], (info)=>{ return { width : 300, height: 300, "transform" : [{"filter": "datum.kind=='" + info + "'"}], "layer" : [ { "mark": {"type": "point", "size" : 150, clip: true, filled : true}, "title" : info, "encoding": { "x": { field: "sites", type: "quantitative", axis: {"title": "Sites", titleFontSize : 14, grid : false}, "scale": {"type": "log"} }, "y": {"field": "value", "type": "quantitative", axis: { "title": "Variants", "titleFontSize" : 14, "grid" : false, }, "scale": { "type": "log", }}, "color" : { "sort": {"field": "gene"}, "field" : "gene" , "type": "nominal", "scale": {"scheme": "set1"}, "legend" : {"title" : "Gene", titleFontSize : 14, labelFontSize: 12, orient : "right"} } } }, { "mark": { "type": "line", "color": "black" }, "transform": [ { "regression": "value", "on": "sites", //"method" : "poly" } ], "encoding": { "x": { "field": "sites", "type": "quantitative" }, "y": { "field": "value", "type": "quantitative" } } }, { "transform": [ { "regression": "value", "on": "sites", //"method" : "poly", "params": true }, {"calculate": "'R²: '+format(datum.rSquared, '.2f')", "as": "R2"} ], "mark": { "type": "text", "color": "firebrick", "x": "width", "align": "right", "y": -5 }, "encoding": { "text": {"type": "nominal", "field": "R2"} } } ] } }) }) 34:Tabb,counts_scatter_hla = vegalite({ "title": null, data: {values: by_gene[2]}, "hconcat" : [ { width : 300, height: 300, "mark": {"type": "point", "size" : 150, clip: true, filled : true}, "title" : null, "encoding": { "x": { field: "positive_in_hla", type: "quantitative", axis: {"title": "Positively Selected Sites in HLA epitopes", titleFontSize : 14, grid : false}, "scale": { "type": "linear", "domain" : [-0.01, 0.15] } }, "y": {"field": "positive_out_hla", "type": "quantitative", axis: { "title": "Positively Selected Sites NOT in HLA epitopes", "titleFontSize" : 14, "grid" : false, }, "scale": { "type": "linear", "domain" : [-0.01, 0.15] }}, "color" : { "sort": {"field": "gene"}, "field" : "gene" , "type": "nominal", "scale": {"scheme": "set1"}, "legend" : {"title" : "Gene", titleFontSize : 14, labelFontSize: 12, orient : "right"} } } }, { width : 300, height: 300, "mark": {"type": "point", "size" : 150, clip: true, filled : true}, "title" : null, "encoding": { "x": { field: "negative_in_hla", type: "quantitative", axis: {"title": "Negatively Selected Sites in HLA epitopes", titleFontSize : 14, grid : false}, "scale": {"type": "linear","domain" : [-0.01, 0.15]} }, "y": {"field": "negative_out_hla", "type": "quantitative", axis: { "title": "Negatively Selected Sites NOT in HLA epitopes", "titleFontSize" : 14, "grid" : false, }, "scale": { "type": "linear", "domain" : [-0.01, 0.15] }}, "color" : { "sort": {"field": "gene"}, "field" : "gene" , "type": "nominal", "scale": {"scheme": "set1"}, "legend" : {"title" : "Gene", titleFontSize : 14, labelFontSize: 12, orient : "right"} } } } ] })35:T51d,aa_variation_plot = vegalite({ "title": null, data: {values: temporal_summary}, transform: [{"calculate" : "datum.prot_num_var_sites_minor/datum.num_sites", "as" : "fraction_2aa"}, {"calculate" : "datum.codon_num_var_sites/datum.num_sites", "as" : "fraction_nuc"}], hconcat : _.map ([["Fraction of sites with shared amino-acid changes", "fraction_2aa"], ["Fraction of sites with any changes", "fraction_nuc"]], (r)=> { return { width : 300, height: 300, "mark": {"type": "line", "interpolate" : "monotone", "strokeWidth" : 3}, "encoding": { "x": {field: "analysis_date", type: "temporal", axis: {"title": "Analysis date", titleFontSize : 14, grid : false}}, "y": {"field": r[1], "type": "quantitative", axis: { "title": r[0], "titleFontSize" : 14, "grid" : false }, "scale": { "type": "linear" }}, "color" : { "sort": {"field": "gene"}, "field" : "gene" , "type": "nominal", "scale": {"scheme": "set1"}, "legend" : {"title" : "Gene", titleFontSize : 14, labelFontSize: 12, orient : "right"} }} }}) })36:T438,bl_plot = vegalite({ "title": null, data: {values: temporal_summary}, hconcat : _.map ([["Total branch length leaves", "tree_len_terminal"], ["Total branch length internal", "tree_len_internal"]], (r)=> { return { width : 300, height: 300, "mark": {"type": "line", "interpolate" : "monotone", "strokeWidth" : 3}, "encoding": { "x": {field: "analysis_date", type: "temporal", axis: {"title": "Analysis date", titleFontSize : 14, grid : false}}, "y": {"field": r[1], "type": "quantitative", axis: { "title": r[0], "titleFontSize" : 14, "grid" : false }, "scale": { "type": "linear", }}, "color" : { "sort": {"field": "gene"}, "field" : "gene" , "type": "nominal", "scale": {"scheme": "set1"}, "legend" : {"title" : "Gene", titleFontSize : 14, labelFontSize: 12, orient : "right"} }} }}) })37:T46e,omega_variation_plot = vegalite({ "title": null, data: {values: temporal_summary}, hconcat : _.map ([["dN/dS Leaves", "mean_dnds_leaf"], ["dN/dS Internal Branches", "mean_dnds_internal"]], (r)=> { return { width : 300, height: 300, "mark": {"type": "line", "interpolate" : "monotone", "strokeWidth" : 3, clip: true}, "encoding": { "x": {field: "analysis_date", type: "temporal", axis: {"title": "Analysis date", titleFontSize : 14, grid : false}}, "y": {"field": r[1], "type": "quantitative", axis: { "title": r[0], "titleFontSize" : 14, "grid" : false, }, "scale": { "type": "linear", "domain" : [0,2], }}, "color" : { "sort": {"field": "gene"}, "field" : "gene" , "type": "nominal", "scale": {"scheme": "set1"}, "legend" : {"title" : "Gene", titleFontSize : 14, labelFontSize: 12, orient : "right"} }} }}) })38:T5ac,temporal_S = { return _.map ([["S","Fraction of sites",sr_plot,"sqrt"],["S"," ",sr_plot_2,"linear"],["S"," ",sr_plot_3,"linear"]], (r)=>vegalite({ "title": null, data: {values: r[2]}, width : 300, height: 300, //"title" : r[0], "transform": [ { "filter": "datum.gene == '" + r[0] + "'" } ], "mark": {"type": "line", "interpolate" : "monotone", "strokeWidth" : 3, tooltip: true}, "encoding": { "x": {field: "date", type: "temporal", axis: {"title": "Analysis date", titleFontSize : 14, grid : false}}, "y": {"field": "value", "type": "quantitative", axis: { "title": r[1], "titleFontSize" : 14, "grid" : false, }, "scale": { "type": r[3], }}, "color" : { "sort": {"field": "type"}, "field" : "type" , "type": "nominal", "scale": {"scheme": "dark2"}, "legend" : {"title" : "", symbolSize: 300, symbolStrokeWidth: 3, titleFontSize : 14, labelLimit: 200, labelFontSize: 14, orient : "right", } }} }))}39:T73a,vegalite({ "title": null, data: {values: site_tracking}, width : 300, height: 600, encoding: { "color" : { "sort": {"field": "nuc"}, "field" : "coordinate" , "type": "nominal", "scale": {"scheme": "category20b"}, "legend" : null }, }, //"title" : r[0], layer: [{ "mark": {"type": "line", "interpolate" : "monotone", "strokeWidth" : 3, "tooltip" : true}, "encoding": { "x": {field: "date", type: "temporal", axis: {"title": "Analysis date", titleFontSize : 14, grid : false, "labelFontSize" : 12}}, "y": {"field": "p", "type": "quantitative", axis: { "title": "MEME p-value", "titleFontSize" : 14, "labelFontSize" : 12, "grid" : false, }, "scale": { "type": "pow", "exponent" : 0.25 }}, "size" : { "condition": {"test": "datum['coordinate'] == 'S 614' || datum['coordinate'] == 'RdRp 323' || datum['coordinate'] == 'S 936' || datum['coordinate'] == 'S 943'", "value": "3"}, "value" : 1.5, } } }, {"encoding": { "x": {"aggregate": "max", "field": "date", "type": "temporal"}, "y": {"aggregate": {"argmax": "date"}, "field": "p", "type": "quantitative"}, }, "layer": [{ "mark": {"type": "circle"} }, { "mark": {"type": "text", "align": "left", "dx": 10, "fontSize" : 12}, "encoding": { "text": {"field":"coordinate", "type": "nominal"}, } }]} ] })3a:T82c,omega_vs_freq = vegalite({ "title": null, data: {values: omega_freq ("")}, layer: [ { width : 400, height: 400, "mark": {"type": "bar", clip: true, filled : true, color: 'grey', opacity : 0.5}, "encoding": { "x": { field: "count", type: "ordinal", axis: {"title": "Variant Frequency", titleFontSize : 20, grid : false, labelFontSize: 14 }, "scale" : {"type": "ordinal", "domain" : ["1","2","3","4","5","6","7","8","9","10","11-20","21-100","101-3000",">3000"]} }, "y": {"field": "number", "type": "quantitative", axis: { "orient": "right", "title": "Count", "titleFontSize" : 20, labelFontSize: 14, titleColor : "grey", "grid" : false, }, "scale": { "type": "sqrt", }}, } }, { width : 400, height: 400, "mark": {"type": "point", "size" : 100, clip: true, filled : true, "color": "black", "tooltip": {"content": "data"}}, "encoding": { "x": { field: "count", type: "ordinal", axis: {"title": "Variant Frequency", titleFontSize : 20, grid : false, labelFontSize: 14 }, "scale" : {"type": "ordinal", "domain" : ["1","2","3","4","5","6","7","8","9","10","11-20","21-100","101-3000",">3000"]} }, "y": {"field": "omega", "type": "quantitative", axis: { "title": "dN/dS", "titleFontSize" : 20, labelFontSize: 14, "grid" : false, }, "scale": { "type": "linear", "domain" : [0,2] }}, } }, ], "resolve": {"scale": {"y": "independent"}} })3b:T7bc,omega_scan = vegalite({ "title": null, data: {values: _.map (_.range (0,27501,250), (d)=>{let counts = _.filter (omega_freq ("",[d,d+2500]), (f)=>f.count == "6")[0]; return {'x' : d, 'omega' : counts.omega, 'number' : counts.number}})}, layer: [ { width : 800, height: 400, "mark": {"type": "line", clip: true, filled : false, color: 'grey', opacity : 1.0, tooltip : true, "interpolate" : "basis"}, "encoding": { "x": { field: "x", type: "quantitative", axis: {"title": "Genomic Coordinate", titleFontSize : 20, grid : false, labelFontSize: 14 }, }, "y": {"field": "omega", "type": "quantitative", axis: { "orient": "left", "title": "dN/dS", "titleFontSize" : 20, labelFontSize: 14, titleColor : "grey", "grid" : false, }, "scale": { "type": "linear", }}, } }, { width : 800, height: 400, "mark": {"type": "line", clip: true, filled : false, color: 'pink', opacity : 1.0, tooltip : true, "interpolate" : "basis"}, "encoding": { "x": { field: "x", type: "quantitative", }, "y": {"field": "number", "type": "quantitative", axis: { "orient": "right", "title": "Count", "titleFontSize" : 20, labelFontSize: 14, titleColor : "pink", "grid" : false, }, "scale": { "type": "linear", }}, } } ], "resolve": {"scale": {"y": "independent"}} })3c:T54a,/*ag_omega_vs_freq = vegalite({ "title": null, data: {values: codons_by_counts}, "encoding": { //"size" : {"value" : 100}, "x": { field: "AF", "title" : "Variant Frequency", type: "ordinal", "scale" : {"type": "ordinal", "domain" : ["1","2","3","4","5","6","7","8","9","10","11-20","21-100","101-3000",">3000"]} }, "y": {"field": "count", "type": "quantitative", "stack": "normalize", axis: { "title" : "fraction", axis: null, "grid" : false, }, "scale": { "type": "linear", }}, "color" : { "legend" : null, "field" : "codon" , "type": "nominal", "scale":{"scheme": "category20c"} }, }, layer: [{ width : 400, height: 1400, "mark": { "type": "bar", "filled" : false, clip : true, tooltip: true, }, }, { "mark": { "type": "text", "align": "center", "baseline": "top", //"dx": 10, }, "encoding": { "text": {"field": "codon", "type": "nominal"} }}] })*/3d:Td88,by_gene = { let by_gene = {}; _.each (comparative_data, (site)=> { if (!(site["G"] in by_gene)) { by_gene[site["G"]] = { 'gene' : site["G"], 'sequences' : d3.sum (_.values(site['cdn'])), 'variable' : 0, 'total' : 0, 'variable-aa' : 0, 'variable-2x' : 0, 'variable-2x-aa' : 0, 'selected-positive' : 0, 'selected-negative' : 0, 'in_epitope' : 0, 'in_epitope_positive' : 0, 'in_epitope_negative' : 0 }; } let ns = false; by_gene[site["G"]]['total'] ++; let cdn_variants = _.sortBy (_.filter (site['cdn'], (v,k)=> {return k.indexOf ('-') < 0 && k != 'NNN'})); if (cdn_variants.length > 1) { by_gene[site["G"]]['variable'] ++; if (_.some (cdn_variants, (v,i) => {return v>=2&&i {return k != '-'})); if (cdn_variants.length > 1) { by_gene[site["G"]]['variable-aa'] ++; if (_.some (cdn_variants, (v,i) => {return v>=2&&ig.sequences > 0), (d)=>-d["selected-positive"]), (v)=>{ _.each (['variable-aa', 'variable', 'variable-2x', 'variable-2x-aa', 'selected-positive', 'selected-negative', 'in_epitope', 'in_epitope_positive', 'in_epitope_negative'], (k)=>{v[k] = [v[k],(v[k]/v["total"])*1000/3]}); return v; }); return [ gene_data, _.flatten (_.map (gene_data, (r)=> { return _.map (['variable-aa','variable','variable-2x','variable-2x-aa','selected-positive','selected-negative'], (m)=>{return {"gene" : r['gene'],'sites': r['total'],'kind' : m, 'value' : Math.max (r[m][0],1)};})} )), _.flatten (_.map (gene_data, (r)=> { return {"gene" : r['gene'], 'positive_in_hla' : r['in_epitope_positive'][0] / r['in_epitope'][0], 'negative_in_hla' : r['in_epitope_negative'][0] / r['in_epitope'][0], 'positive_out_hla': (r['selected-positive'][0]-r['in_epitope_positive'][0]) / (r['total']-r['in_epitope'][0]), 'negative_out_hla': (r['selected-negative'][0]-r['in_epitope_negative'][0]) / (r['total']-r['in_epitope'][0]), }; })) ]; }3e:T47b,by_gene_comp = { let by_gene = {}; _.each (comparative_data, (site)=> { if (!(site["G"] in by_gene)) { by_gene[site["G"]] = { 'gene' : site["G"], 'selected-positive' : 0, 'selected-negative' : 0, 'different+' : 0, 'different-' : 0 }; } let dp = false; if ("bFEL" in site) { if (site.bFEL.p <= pv) { if (site.bFEL.a < site.bFEL.b) { by_gene[site["G"]]['selected-positive'] ++; dp = true; } else { by_gene[site["G"]]['selected-negative'] ++; } } } if (("bMEME" in site) && !dp) { if (site.bMEME.p <= pv) { by_gene[site["G"]]['selected-positive'] ++; } } if (("bCFEL" in site)) { if (site.bCFEL.p <= pv) { if (site.bCFEL['b-nCOV'] > site.bCFEL['b']) { by_gene[site["G"]]['different+'] ++; } else { by_gene[site["G"]]['different-'] ++; } } } }); return _.sortBy (by_gene, "gene"); }3f:Td60,by_gene_plot = { let variants = []; _.each (comparative_data, (site, key)=> { let plot_object = { 'gene' : site["G"], 'coord' : +site["S"], 'type' : null, 'kind' : null }; function add_point (kind, type) { plot_object['type'] = type || "Variable"; plot_object['kind'] = kind; variants.push (_.clone ( plot_object)); }; let cdn_variants = _.sortBy (_.filter (site['cdn'], (v,k)=> {return k.indexOf ('-') < 0 && k != 'NNN'})); if (cdn_variants.length > 1) { add_point ("Nucleotide"); if (_.some (cdn_variants, (v,i) => {return v>=2&&i {return k != '-'})); if (aa_variants.length > 1) { add_point ("AA SNP"); if (_.some (aa_variants, (v,i) => {return v>=2&&i{ if (c in intra_host) { ihv = _.extend (ihv, intra_host[c]); } }); _.each (ihv, (v,k)=> { add_point ("Intrahost " + k,"Intrahost variants"); }); if (site["evo"]) { let nuc = {'A': 1, 'C': 1, 'G' :1, 'T': 1}; cdn_variants = _.filter (_.map (site['cdn'], (c,k)=>[k,c]), (c)=>_.every (c[0], (n)=>{return (n in nuc);})); if (_.some (cdn_variants, (c)=> { if (c[0] in site["evo"]) { return false; } if (site["G"] == "S") console.log (c, site); return c[1] >= 5; })) { add_point ("Unexpected codons","Adaptive"); } } }); return variants; }40:T947,omega_freq = (gene, coordinates) => { /*GENE: "ORF1a" SITE: 265 CODON: "TTG" REFERENCE: "ATG" AF: 1 S: 0 N: 1 ES: 0.007465336396128932 EN: 2.999542051752158*/ let exclude_codons = new Set (excuded_codons); gene = gene || ""; coordinates = coordinates || [0,50000]; let raw_data = _.filter (raw_slac_data, (d)=> (gene.length == 0 || gene == d.GENE) && (d.SITE >= coordinates [0] && d.SITE <= coordinates[1]) && !exclude_codons.has (d.CODON )); return _.filter (_.flatten(_.map (_.groupBy (raw_data, (d)=> { if (d.AF <= 10) { return d.AF; } if (d.AF <= 20) { return "11-20"; } if (d.AF <= 100) { return "21-100"; } if (d.AF <= 3000) { return "101-3000"; } return ">3000"; }), (codon,AF)=> { let uniq_sites = _.uniqBy (codon, (cdn)=>cdn.SITE); let ES = d3.sum (uniq_sites, (cnt)=>(cnt.ES <= 3 ? cnt.ES : 0.)); let EN = d3.sum (uniq_sites, (cnt)=>(cnt.EN <= 3 ? cnt.EN : 0.)); let N = d3.sum (codon, (cnt)=>cnt.N); let S = d3.sum (codon, (cnt)=>cnt.S); return { 'count' : AF, 'number' : codon.length, 'N' : N, 'S' : S, 'ES' : ES, 'EN' : EN, 'omega' : (N/S)/(EN/ES) }; } /* let ES = d3.sum (d, v=>{ if (! (k in sites_already_counted)) { sites_already_counted[k] = {}; } if (v.SITE in sites_already_counted[k]) { return 0; } sites_already_counted[k] = {}; sites_already_counted[k][v.SITE] = 1; return v.ES; }); let EN = d3.sum (d, v=>{ if (! (k in ns_sites_already_counted)) { ns_sites_already_counted[k] = {}; } if (v.SITE in ns_sites_already_counted[k]) { return 0; } ns_sites_already_counted[k] = {}; ns_sites_already_counted[k][v.SITE] = 1; return v.EN; }); let S = d3.sum (d, v=>v.S); let N = d3.sum (d, v=>v.N); let count = d.length; return {'count' : k, 'number' : count, 'ES' : ES, 'EN' : EN, 'N' : N, 'S' : S, 'omega' : (N/S)/(EN/ES)}; */ )), (e)=>true); }41:T677,aggregated_omega_freq = { let resolved = {'A' : 1, 'C': 1, 'G' : 1, 'T' : 1}; let bins = {}; _.each (_.filter (comparative_data, (d)=> ('SLAC' in d) && (d.SLAC.S + d.SLAC.N)), (d)=> { let passing_filter = _.filter (_.map (d.cdn, (count, codon) => [codon, count]), (cdn) => { return _.every (cdn[0], (letter)=>(letter in resolved)); }); let weight = d3.sum (passing_filter, (f) => f[1]); if (_.filter (passing_filter, (f)=>f[0] != d.bSC2).length > 1) { return; } let all_variants = d3.sum (_.filter (passing_filter, (f)=>f[0] != d.bSC2), (f) => f[1]); let rv = all_variants; if (all_variants > 6) { if (all_variants <= 10) { all_variants = "7-10" } else { if (all_variants <= 20) { all_variants = "11-20" } else { if (all_variants <= 3000) { all_variants = "21-3000"; } else { all_variants = ">3000"; } } } } else { all_variants = "" + all_variants; } if (!(all_variants in bins)) { bins[all_variants] = {'N' : 0, 'S' : 0, 'EN' : 0, 'ES' : 0, 'count' : 0, 'vars': []}; } bins[all_variants]['S'] += d.SLAC.S; bins[all_variants]['N'] += d.SLAC.N; bins[all_variants]['ES'] += d.SLAC.ES; bins[all_variants]['EN'] += d.SLAC.EN; bins[all_variants]['vars'].push (passing_filter); bins[all_variants]['count'] ++; }); return bins; }42:T50c,sr_plot = _.flatten(_.map (_.filter (temporal_summary, (r)=>true), (r)=>{ let mapper = [ //["Unique haplotypes", (d)=>(+d.num_uniq_haps)/(+d.num_seqs)], ["Nucleotide variants" , (d)=>(+d.codon_num_var_sites)/(+d.num_sites)], ["Shared nucleotide variants" , (d)=>(+d.codon_num_var_sites_minor)/(+d.num_sites)], ["Aminoacid variants" , (d)=>(+d.prot_num_var_sites)/(+d.num_sites)], ["Shared aminoacid variants" , (d)=>(+d.prot_num_var_sites_minor)/(+d.num_sites)], ["Positively selected", (d)=>(+d.num_sites_meme)/(+d.num_sites)] ]; /* analysis_date: 2020-03-30T00:00 gene: "ORF7a" num_seqs: "2473" num_uniq_haps: "35" VPH: "1.0" VPH_Q_1: "1.0" VPH_Q_3: "1.5" tree_len_internal: 0.002871102987799821 tree_len_terminal: "0.08568267397569826" mean_dnds_leaf: "1.067260875707685" mean_dnds_internal: "0" num_sites: "121" codon_num_var_sites: "32" codon_num_var_sites_minor: "1" prot_num_var_sites: "27" prot_num_var_sites_minor: "0" num_sites_pos_fel: "0" num_sites_neg_fel: "0" num_sites_meme: "0" median_branches_meme: "nan" */ return _.map (mapper, (tf)=> { let tp = (tf[1])(r); return { 'date' : r.analysis_date, 'gene' : r.gene, "type" : tf[0], "value" : tp } }); }))43:T41c,sr_plot_3 = _.flatten(_.map (_.filter (temporal_summary, (r)=>true), (r)=>{ let mapper = [ //["Unique haplotypes", (d)=>(+d.num_uniq_haps)/(+d.num_seqs)], ["Total internal branch length", (d)=>+d.tree_len_internal], ["dN/dS (internal)", (d)=>+d.mean_dnds_internal], ["dN/dS (leaves)", (d)=>+d.mean_dnds_leaf] ]; /* analysis_date: 2020-03-30T00:00 gene: "ORF7a" num_seqs: "2473" num_uniq_haps: "35" VPH: "1.0" VPH_Q_1: "1.0" VPH_Q_3: "1.5" tree_len_internal: 0.002871102987799821 tree_len_terminal: "0.08568267397569826" mean_dnds_leaf: "1.067260875707685" mean_dnds_internal: "0" num_sites: "121" codon_num_var_sites: "32" codon_num_var_sites_minor: "1" prot_num_var_sites: "27" prot_num_var_sites_minor: "0" num_sites_pos_fel: "0" num_sites_neg_fel: "0" num_sites_meme: "0" median_branches_meme: "nan" */ return _.map (mapper, (tf)=> { let tp = (tf[1])(r); return { 'date' : r.analysis_date, 'gene' : r.gene, "type" : tf[0], "value" : tp } }); }))44:[] 2f:{"title":"Summary of SARS-CoV-2 genomic diversity","cells":[{"id":0,"value":"md`# Summary of SARS-CoV-2 genomic diversity`","pinned":false,"hidden":false,"mode":"ojs"},{"id":784,"value":"md`A quick summary of variation found in SARS-CoV-2 genomic samples from GISAID. Main analysis page is [here](https://observablehq.com/@spond/natural-selection-analysis-of-sars-cov-2-covid-19)`","pinned":false,"hidden":false,"mode":"ojs"},{"id":963,"value":"md`### Summary of sequence diversity/selection in SARS-CoV-2 isolates by gene/ORF`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1345,"value":"viewof per_kp = checkbox({\n options: [{ value: \"off\", label: \"Show counts per kB\" }],\n value: \"toggle\",\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":943,"value":"$31","pinned":false,"hidden":false,"mode":"ojs"},{"id":966,"value":"md`### Summary of sequence diversity/selection in related beta-CoV by gene/ORF`","pinned":false,"hidden":false,"mode":"ojs"},{"id":978,"value":"table (by_gene_comp,\n {\n limit: 5000,\n enableCSVDownload: true,\n sort: true,\n enableFilter: false,\n columns: [\n \n {\n key: 'gene',\n name: 'Gene/ORF'\n },\n {\n key: 'selected-positive',\n name: \"+ selection\",\n },\n {\n key: 'selected-negative',\n name: \"- selection\",\n },\n {\n key: 'different+',\n name: \"Less conservation in nCoV\",\n },\n {\n key: 'different-',\n name: \"More conservation in nCoV\",\n }\n ]\n});","pinned":false,"hidden":false,"mode":"ojs"},{"id":887,"value":"html`\n \n

`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1233,"value":"viewof gene = {\n const dd3 = select({\n title: \"View this gene\",\n options: _.map (_.uniqBy (by_gene[0],\"gene\"), (g)=>g.gene)\n \n });\n dd3.input.style.fontSize = \"30px\";\n dd3.input.style.marginTop = \"8px\";\n return dd3;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":996,"value":"$32","pinned":false,"hidden":false,"mode":"ojs"},{"id":1487,"value":"md`### Relationship between gene/ORF length and the number of variable sites`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1354,"value":"$33","pinned":false,"hidden":false,"mode":"ojs"},{"id":1503,"value":"md`### Relationship between gene/ORF sites in HLA epitopes and the number of sites selected`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1499,"value":"$34","pinned":false,"hidden":false,"mode":"ojs"},{"id":1540,"value":"md`### Temporal trends in per-site variability in different genes`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1230,"value":"$35","pinned":false,"hidden":false,"mode":"ojs"},{"id":1543,"value":"md`### Temporal trends in cumulative branch lengths`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1326,"value":"$36","pinned":false,"hidden":false,"mode":"ojs"},{"id":1548,"value":"md`### dN/dS temporal trends`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1264,"value":"$37","pinned":false,"hidden":false,"mode":"ojs"},{"id":1272,"value":"omega_scatter = vegalite({\n \"title\": null,\n data: {values: temporal_summary},\n\n width : 300,\n height: 300,\n\n \"mark\": {\"type\": \"point\", \"strokeWidth\" : 3, clip: true},\n \"encoding\": {\n \"x\": {field: \"mean_dnds_leaf\", type: \"quantitative\", axis: {\"title\": \"dN/dS terminal braches\", titleFontSize : 14, grid : false}},\n\n \"y\": {\"field\": \"mean_dnds_internal\", \"type\": \"quantitative\", axis: {\n \"title\": \"dN/dS internal braches\",\n \"titleFontSize\" : 14,\n \"grid\" : false,\n \n },\n \"scale\": { \n \"type\": \"linear\",\n \"domain\" : [0,2],\n }},\n \"color\" : {\n \"sort\": {\"field\": \"gene\"},\n \"field\" : \"gene\" ,\n \"type\": \"nominal\",\n \"scale\": {\"scheme\": \"set1\"},\n \"legend\" : {\"title\" : \"Gene\", titleFontSize : 14, labelFontSize: 12, orient : \"right\"}\n }}\n\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":2351,"value":"temporal_S[0]","pinned":false,"hidden":false,"mode":"ojs"},{"id":2355,"value":"temporal_S[1]","pinned":false,"hidden":false,"mode":"ojs"},{"id":2362,"value":"temporal_S[2]","pinned":false,"hidden":false,"mode":"ojs"},{"id":2221,"value":"$38","pinned":false,"hidden":false,"mode":"ojs"},{"id":2386,"value":"$39","pinned":false,"hidden":false,"mode":"ojs"},{"id":1551,"value":"md`### Distributions of variant frequencies`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1285,"value":"variant_histograms = vegalite({\n \"title\": \"Amino-acid variants at ≤1.0%\",\n width : 400,\n height: 400,\n data: {values: variant_frequencies},\n transform :[{\"filter\" : \"datum[4]<0.01\"}],\n \"mark\": {\"type\": \"bar\"},\n\n encoding: {\n /*\"facet\" : {\n \"field\" : \"2\" ,\n \"type\": \"ordinal\",\n \"columns\": 3,\n \"title\" : \"Amino-acid\",\n \"sort\": {\"field\": \"2\"}\n },*/\n \n \"x\": {field: \"4\", \n type: \"quantitative\", \n axis: {\"title\": \"Variant frequency\", titleFontSize : 14, grid : false, format: \"2p\"}, \n \"bin\": {\n \"bin\" :true,\n \"maxbins\" : 50\n }\n \n },\n \"y\": {\n axis: {\"title\": \"Count\", titleFontSize : 14, grid : false},\n \"aggregate\": \"count\",\n \"type\": \"quantitative\"\n }\n }\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":1322,"value":"variant_histograms_common = vegalite({\n \"title\": \"Amino-acid variants at >1.0%\",\n width : 400,\n height: 400,\n data: {values: variant_frequencies},\n transform :[{\"filter\" : \"datum[4]>0.01\"}],\n \"mark\": {\"type\": \"bar\"},\n\n encoding: {\n /*\"facet\" : {\n \"field\" : \"2\" ,\n \"type\": \"ordinal\",\n \"columns\": 3,\n \"title\" : \"Amino-acid\",\n \"sort\": {\"field\": \"2\"}\n },*/\n \n \"x\": {field: \"4\", \n type: \"quantitative\", \n axis: {\"title\": \"Variant frequency\", titleFontSize : 14, grid : false, format: \"2p\"}, \n \"bin\": {\n \"bin\" :true,\n \"maxbins\" : 50\n }\n \n },\n \"y\": {\n axis: {\"title\": \"Count\", titleFontSize : 14, grid : false},\n \"aggregate\": \"count\",\n \"type\": \"quantitative\"\n }\n }\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":2081,"value":"_.sortBy (_.map (_.groupBy (_.filter (raw_slac_data, (d)=>d.AF == 6), (d)=>d.CODON), (d,k)=>[k,d3.sum (d, d=>d.N-d.S)]), d=>-d[1])","pinned":false,"hidden":false,"mode":"ojs"},{"id":2040,"value":"viewof excuded_codons = checkbox({\n description: \"Exclude these codons\",\n options: _.sortBy (_.map (_.filter (translation_table, (t)=>t.AA != \"*\"), (c)=>c.Codon)),\n multiple: true\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":1574,"value":"$3a","pinned":false,"hidden":false,"mode":"ojs"},{"id":2130,"value":"$3b","pinned":false,"hidden":false,"mode":"ojs"},{"id":1771,"value":"$3c","pinned":false,"hidden":false,"mode":"ojs"},{"id":5,"value":"md`---`","pinned":false,"hidden":false,"mode":"ojs"},{"id":930,"value":"pv = 0.1","pinned":false,"hidden":false,"mode":"ojs"},{"id":907,"value":"$3d","pinned":false,"hidden":false,"mode":"ojs"},{"id":1252,"value":"variant_frequencies = {\n let variants = [];\n _.each (comparative_data, (site)=> {\n var denominator = d3.sum (_.values (site['aa']));\n _.each (site['aa'], (v,k)=> {\n if (k != site['bSC2-aa']) {\n if (k.indexOf('-')<0) {\n variants.push ([site['G'],site['S'],k,v,v/denominator,v>1 ? \"Repeated\" : \"Singleton\"]);\n }\n }\n })\n });\n return variants;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":971,"value":"$3e","pinned":false,"hidden":false,"mode":"ojs"},{"id":1006,"value":"$3f","pinned":false,"hidden":false,"mode":"ojs"},{"id":2050,"value":"raw_slac_data = await d3.csv (\"https://dl.dropbox.com/s/g9b45bv5xybg1bx/sequences.by_count.csv\", d3.autoType)","pinned":false,"hidden":false,"mode":"ojs"},{"id":1554,"value":"$40","pinned":false,"hidden":false,"mode":"ojs"},{"id":2185,"value":"omega_freq()","pinned":false,"hidden":false,"mode":"ojs"},{"id":1901,"value":"codons_by_counts = {\n let raw_data = await d3.csv ('https://dl.dropbox.com/s/pyvwwactcyqgxph/sequences.by_count_codons.csv', d3.autoType);\n return _.flatten (_.map (_.groupBy (raw_data, (d)=> {\n if (d.AF <= 10) {\n return d.AF; \n }\n if (d.AF <= 20) {\n return \"11-20\";\n }\n if (d.AF <= 100) {\n return \"21-100\";\n }\n if (d.AF <= 3000) {\n return \"101-3000\";\n }\n return \">3000\";\n }), (codons, AF)=> {\n return _.map (_.groupBy (codons, \"CODON\"), (counts, codon)=> {\n return {'AF' : AF, 'codon' : codon, 'count' : d3.sum (counts, (d)=>d.COUNT)};\n });\n }));\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1712,"value":"$41","pinned":false,"hidden":false,"mode":"ojs"},{"id":904,"value":"comparative_data = await d3.json (\"https://raw.githubusercontent.com/veg/SARS-CoV-2/segmented/data/comparative-annotation.json\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":1155,"value":"temporal_summary = {\n let ts = await d3.csv (\"https://data.hyphy.org/web/covid-19/selection-analyses/report.csv\");\n let parsers = [d3.timeParse (\"%Y-%m-%d\"), d3.timeParse (\"%Y%m%d\")];\n let valid_dates = {};\n \n _.each (ts, (r)=> {\n if (!(r['analysis_date'] in valid_dates)) {\n valid_dates [r['analysis_date']] = {}\n }\n if (_.every (r, (f)=>f.length != 0)) {\n valid_dates [r['analysis_date']][r[\"gene\"]] = 1;\n }\n });\n \n \n ts = _.filter (ts, (t)=>valid_dates[t['analysis_date']]['S'] && valid_dates[t['analysis_date']]['RdRp']);\n _.each (ts, (r)=> {\n r['raw_date'] = r['analysis_date'];\n r['analysis_date'] =_.find (_.map (parsers, (p)=>p (r['analysis_date'])));\n r['tree_len_internal'] = parseFloat (r['tree_len_internal']);\n });\n return ts;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2366,"value":"d3.median (_.sortBy (_.map (_.filter(temporal_summary, (d)=>d.raw_date == \"2020-06-28\"), (d)=>+d.num_seqs)))","pinned":false,"hidden":false,"mode":"ojs"},{"id":2228,"value":"$42","pinned":false,"hidden":false,"mode":"ojs"},{"id":2333,"value":"sr_plot_2 = _.flatten(_.map (_.filter (temporal_summary, (r)=>true), (r)=>{\n let mapper = [\n//[\"Unique haplotypes\", (d)=>(+d.num_uniq_haps)/(+d.num_seqs)],\n [\"Sequences\" , (d)=>+d.num_seqs],\n [\"Unique haplotypes\" , (d)=>+d.num_uniq_haps],\n ];\n \n /*\n analysis_date: 2020-03-30T00:00\n gene: \"ORF7a\"\n num_seqs: \"2473\"\n num_uniq_haps: \"35\"\n VPH: \"1.0\"\n VPH_Q_1: \"1.0\"\n VPH_Q_3: \"1.5\"\n tree_len_internal: 0.002871102987799821\n tree_len_terminal: \"0.08568267397569826\"\n mean_dnds_leaf: \"1.067260875707685\"\n mean_dnds_internal: \"0\"\n num_sites: \"121\"\n codon_num_var_sites: \"32\"\n codon_num_var_sites_minor: \"1\"\n prot_num_var_sites: \"27\"\n prot_num_var_sites_minor: \"0\"\n num_sites_pos_fel: \"0\"\n num_sites_neg_fel: \"0\"\n num_sites_meme: \"0\"\n median_branches_meme: \"nan\"\n */\n \n return _.map (mapper, (tf)=> {\n let tp = (tf[1])(r);\n return {\n 'date' : r.analysis_date, 'gene' : r.gene, \"type\" : tf[0], \"value\" : tp\n }\n });\n}))","pinned":false,"hidden":false,"mode":"ojs"},{"id":2341,"value":"$43","pinned":false,"hidden":false,"mode":"ojs"},{"id":1089,"value":"gene_coordinates = _.map ([\n [266,'ORF1a',13483],\n [13471, 'ORF1b',21555],\n [21563,'S',25385],\n [25393,'ORF3a',26220],\n [26523,'M',27191],\n [27202,'ORF6',27387],\n [27394,'ORF7a',27759],\n [27894,'ORF8',28259],\n [28274,'N',29533],\n //[29558,'ORF10',29674]\n], (d,i)=>{\n d.push (i);\n return d;\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":1087,"value":"intra_host = {\n let raw_variants = await d3.tsv (\"https://raw.githubusercontent.com/galaxyproject/SARS-CoV-2/master/genomics/4-Variation/variant_list.05.tsv\", d3.autoType);\n /*\n Sample\tCHROM\tPOS\tREF\tALT\tDP\tAF\tSB\tDP4\tIMPACT\tFUNCLASS\tEFFECT\tGENE\tCODON\ttype\n \"*/\n let binned = _.countBy (raw_variants, (v)=>v.Sample);\n let vars = {};\n let data = _.each (_.filter (raw_variants, (v)=>v.POS >= gene_coordinates[0][0] && binned[v.Sample] <= 75 && v.AF >= 0.05 && v.AF <= .90), (v)=> {\n let gene = _.find (gene_coordinates, (d,i)=> i == gene_coordinates.length -1 || gene_coordinates[i][0] <= v.POS && gene_coordinates[i+1][0] > v.POS);\n \n let offset = 0;\n if (gene[1] == \"ORF1b\") {\n offset = 3; \n }\n if (!(v.POS in vars)) {\n vars[v.POS] = {}; \n }\n vars[v.POS][v['EFF[*].FUNCLASS']] = true;\n \n });\n \n \n\n return vars;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":9,"value":"d3 = require (\"d3\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":13,"value":"_=require(\"lodash\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":30,"value":"vegalite = require(\"vega-embed\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":764,"value":"import {select, slider, checkbox} from \"@jashkenas/inputs\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":841,"value":"import {rangeSlider} from \"@mootari/range-slider\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":940,"value":"import { table } from \"@gampleman/table\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":2059,"value":"translation_table = {\n var code = d3.csvParse (\"Codon,AA\\nTTT,F\\nTCT,S\\nTAT,Y\\nTGT,C\\nTTC,F\\nTCC,S\\nTAC,Y\\nTGC,C\\nTTA,L\\nTCA,S\\nTAA,*\\nTGA,*\\nTTG,L\\nTCG,S\\nTAG,*\\nTGG,W\\nCTT,L\\nCCT,P\\nCAT,H\\nCGT,R\\nCTC,L\\nCCC,P\\nCAC,H\\nCGC,R\\nCTA,L\\nCCA,P\\nCAA,Q\\nCGA,R\\nCTG,L\\nCCG,P\\nCAG,Q\\nCGG,R\\nATT,I\\nACT,T\\nAAT,N\\nAGT,S\\nATC,I\\nACC,T\\nAAC,N\\nAGC,S\\nATA,I\\nACA,T\\nAAA,K\\nAGA,R\\nATG,M\\nACG,T\\nAAG,K\\nAGG,R\\nGTT,V\\nGCT,A\\nGAT,D\\nGGT,G\\nGTC,V\\nGCC,A\\nGAC,D\\nGGC,G\\nGTA,V\\nGCA,A\\nGAA,E\\nGGA,G\\nGTG,V\\nGCG,A\\nGAG,E\\nGGG,G\\n\");\n return code;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2375,"value":"site_tracking = {\n return _.map (await d3.csv (\"https://dl.dropbox.com/s/q4js2mbpm0vke9b/sites.csv\", d3.autoType), (d)=> {\n d.nuc = d.coordinate;\n if (d.coordinate > 7000) {\n d.coordinate = \"S \" + (d.coordinate - 7186); \n } else {\n d.coordinate = \"RdRp \" + (d.coordinate - 4482); \n }\n return d;\n });\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2492,"value":"ext=d3.extent (_.map (_.filter (sr_plot_2, (d)=>d.gene == \"S\" && d.type == \"Unique haplotypes\"), (d)=>d.value))","pinned":true,"hidden":false,"mode":"ojs"},{"id":2501,"value":"ext[1]/ext[0]","pinned":false,"hidden":false,"mode":"ojs"}],"files":"$W44","stdlib":"1","titling":"auto"}