2a:T5cd,vegalite({ "data": {"values": _.map (sim_data, (s)=>{return {'delta' : s.delta, 'psi' : s.psi};})}, "hconcat" : [{ "mark": "bar", "encoding": { "x": { "bin" : { "binned" : true, "step" : 0.05 }, "field": "delta", "type": "quantitative", "axis":{ "title" : "Two-hit rate", grid : false, titleFontSize: 18, labelFontSize: 14, } }, "y": { "aggregate": "count", "type": "quantitative", "axis":{ "title" : "Frequency", grid : false, titleFontSize: 18, labelFontSize: 14, format : "f" } }, } }, { "mark": "bar", "encoding": { "x": { "bin" : { "binned" : true, "step" : 0.5 }, "field": "psi", "type": "quantitative", "axis":{ "title" : "Three-hit rate", grid : false, titleFontSize: 18, labelFontSize: 14, } }, "y": { "aggregate": "count", "type": "quantitative", "axis":{ "title" : "Frequency", grid : false, titleFontSize: 18, labelFontSize: 14, format : "f" } }, } }] })2b:T96c,plot_power_curve = (data, field, step, p, pfield, labels, stype, atype) => { pfield = pfield || "p21"; stype = stype || "log"; atype = atype || "median"; _.each (data, (d)=>{ d.called = d[pfield] <= p; }); return vegalite ({"data": {values : data}, "width" : 500, "height" : 500, "layer" : [ { "mark": { "type" : "line", "strokeWidth" : 5, "interpolate" : "monotone" }, "encoding": { "x": { "field": field[0], "type": "quantitative", "bin" : { "binned" : true, "step" : step }, "axis" : { "title" : labels[0], "titleFontSize" : 14 } }, "y": { "aggregate" : atype, "field": pfield, "type": "quantitative", "scale" : { "type" : stype, }, "axis" : { "title" : labels[1], "titleFontSize" : 14 } }, } }, { "mark": { "type" : "area", "strokeWidth" : 1, "fillOpacity": 0.2, "interpolate" : "monotone" }, "encoding": { "x": { "field": field[0], "type": "quantitative", "bin" : { "binned" : true, "step" : step }, }, "y": { "aggregate" : "q1", "field": pfield, "type": "quantitative" }, "y2": { "aggregate" : "q3", "field": pfield, "type": "quantitative", }, } }, { "mark": { "type" : "circle", "color" : "#f58518" }, "encoding": { "x": { "field": field[0], "type": "quantitative", "bin" : { "binned" : true, "step" : step }, }, "y": { "aggregate" : atype, "field": pfield, "type": "quantitative", "scale" : { "type" : stype, } }, "size": { "aggregate": "count", "type": "quantitative", "legend" : {"title" : "Samples", "titleFontSize": 20, "orient" : "right", "labelFontSize" : 14}, } } } ], "config": { "view": { "stroke": "transparent" } }}) }2c:T402,compute_metrics = (data, accessor, p_crit, label) => { var false_positives = 0, false_negatives = 0, true_positives = 0, true_negatives = 0; _.each (data, (d) => { var p = accessor (d); var l = label (d); if (l) { if (p <= p_crit) { true_positives++; } else { false_negatives++; } } else { if (p <= p_crit) { false_positives++; } else { true_negatives++; } } }); var N = true_positives + false_positives + true_negatives + false_negatives; return { 'p' : p_crit, 'N' : N, 'tp' : true_positives, 'fp' : false_positives, 'tn' : true_negatives, 'fn' : false_negatives, 'fpr' : false_positives / (true_negatives + false_positives), 'power' : true_positives / (true_positives + false_negatives), 'ppv' : true_positives / (true_positives + false_positives) }; };2d:T43f,compute_roc = (data, accessor, step, label) => { var aoc_by_step = [[0,0,0,0]]; var sorted_list = []; var total_positives = 0, total_negatives = 0; _.each (data, (d) => { var p = accessor (d); var l = label (d); if (l) { total_positives ++; } else { total_negatives ++; } sorted_list.push ([p,l]); }); sorted_list.sort ((a,b) => a[0]-b[0]); var index = 0; var true_calls = 0, false_calls = 0; for (var p = step; p <= 1 + (step/2); p+=step) { while (index < sorted_list.length && sorted_list[index][0] <= p) { if (sorted_list[index][1]) { true_calls++; } else { false_calls ++; } index ++; } aoc_by_step.push ([p, true_calls+false_calls,true_calls/total_positives,false_calls/total_negatives, true_calls/(true_calls+false_calls)]); } return _.map (aoc_by_step, (d) => {return {'fp' : d[3], 'power' : d[2], 'ppv' : d[4], "p" : d[0]}}); };2e:T82b,plot_prc = (plot_data, labels) => { var merged_data = []; _.each (plot_data, (d,i) => { _.each (d, (e) => { e.label = labels[i]; merged_data.push (e); }); }); var crticial_values = _.filter (merged_data, (d)=>Math.abs (d.p - 0.01) < 1e-5 || Math.abs (d.p - 0.05) < 1e-5 || Math.abs (d.p - 0.1) < 1e-5); return vegalite ({ "layer" : [{ "data": {values : merged_data}, "mark": { "type": "line", "interpolate": "monotone", "strokeWidth" : 5 }, "width" : 500, "height" : 500, "encoding": { "x": { "field": "power", "axis":{ "title" : "Recall", grid : false, titleFontSize: 18, labelFontSize: 14, format : "g" }, "type": "quantitative" }, "y": { "field": "ppv", "type": "quantitative", "axis":{ "title" : "Precision", grid : false, titleFontSize: 18, labelFontSize: 14, format : "g" } }, "color" : { "field" : "label", "type" : "nominal", "legend" : {"title" : "Simulation", "titleFontSize": 1, "orient" : "right", "labelFontSize" : 14}, } } }, { "data": {values : crticial_values}, "mark": { "type": "point", "size" : 400, "filled" : true }, "encoding": { "x": { "field": "power", "axis":{ "title" : "Recall", grid : false, titleFontSize: 14, labelFontSize: 12, format : "g" }, "type": "quantitative" }, "y": { "field": "ppv", "type": "quantitative", "axis":{ "title" : "Precision", grid : false, titleFontSize: 16, labelFontSize: 12, format : "g" } }, "color" : { "field" : "label", "type" : "nominal" } } }] , "config": { "view": { "stroke": "transparent" } }}); }2f:T4b7,plot_qq = (plot_data, label) => { return vegalite ({"data": {values : plot_data}, "layer" : [{ "mark": { "type": "line", "interpolate": "monotone" }, "width" : 300, "height" : 300, "encoding": { "x": { "field": "expected", "axis":{ "title" : "Theoretical", grid : false, titleFontSize: 14, labelFontSize: 12, format : "g" }, "type": "quantitative" }, "y": { "field": "observed", "type": "quantitative", "axis":{ "title" : "Observed", grid : false, titleFontSize: 16, labelFontSize: 12, format : "g" } } } }, { "mark": { "type": "line", "interpolate": "monotone", "color" : "grey", "strokeDash" : [3,3] }, "width" : 800, "height" : 800, "encoding": { "x": { "field": "expected", "type": "quantitative" }, "y": { "field": "expected", "type": "quantitative", } } } ], "config": { "view": { "stroke": "transparent" } }}); }30:[] 29:{"title":"Multi-hit simulations","cells":[{"id":0,"value":"md`# Multi-hit simulations\n\nExploration of null simulation results for the multiple hit model using empirical simulation templates\n\n`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1145,"value":"md`**Figure 1.a** q-q plot for the expected (dotted line) vs observed p-values (2-hit vs 1-hit) for null (no multiple hits) data` ","pinned":false,"hidden":false,"mode":"ojs"},{"id":384,"value":"plot_qq (_.filter (_.map (compute_roc (_.filter (sim_data, (d)=>d.null), (d)=>d.p21, 0.005, (d)=>false), (d)=>{return {\"expected\" : d.p, \"observed\" : d.fp}}), (d)=>d.expected <= 0.1))","pinned":false,"hidden":false,"mode":"ojs"},{"id":1203,"value":"md`**Figure 1.b** q-q plot for the expected (dotted line) vs observed p-values (3-hit vs 1-hit) for null (no multiple hits) data` ","pinned":false,"hidden":false,"mode":"ojs"},{"id":1201,"value":"plot_qq (_.filter (_.map (compute_roc (_.filter (sim_data, (d)=>d.null), (d)=>d.p31, 0.005, (d)=>false), (d)=>{return {\"expected\" : d.p, \"observed\" : d.fp}}), (d)=>d.expected <= 0.1))","pinned":false,"hidden":false,"mode":"ojs"},{"id":1191,"value":"md`**Figure 1.c** q-q plot for the expected (dotted line) vs observed p-values (3-hit vs 2-hit) for null (no triple hits) data` ","pinned":false,"hidden":false,"mode":"ojs"},{"id":1196,"value":"plot_qq (_.filter (_.map (compute_roc (_.filter (sim_data, (d)=>d.delta > 0 && d.psi == 0.0 && d.psi_syn == 0.0), (d)=>d.p32, 0.005, (d)=>false), (d)=>{return {\"expected\" : d.p, \"observed\" : d.fp}}), (d)=>d.expected <= 0.1))","pinned":false,"hidden":false,"mode":"ojs"},{"id":1208,"value":"md`**Figure 1.d** q-q plot for the expected (dotted line) vs observed p-values (3-hit vs 3-hit syn. islands) for null (only synonymous islands triple hits) data` ","pinned":false,"hidden":false,"mode":"ojs"},{"id":1211,"value":"plot_qq (_.filter (_.map (compute_roc (_.filter (sim_data, (d)=> d.delta == 0 && d.psi == 0.0 && d.psi_syn > 0.0), (d)=>d.p3i3, 0.005, (d)=>false), (d)=>{return {\"expected\" : d.p, \"observed\" : d.fp}}), (d)=>d.expected <= 0.1))","pinned":false,"hidden":false,"mode":"ojs"},{"id":1150,"value":"md`**Figure 2.** Median (Q1-Q3) LRT for the two-hit vs one-hit test as a function of the simulated rate δ > 0.05` ","pinned":false,"hidden":false,"mode":"ojs"},{"id":698,"value":"plot_power_curve (_.filter (sim_data, (d)=> d.delta > 0.05), [\"delta\"], 0.05, 0.05,\"LRT21\", [\"Two-hit rate δ\", \"LRT statistic (2-hit vs 1-hit)\"],\"log\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":1162,"value":"md`**Figure 3.** Median (Q1-Q3) LRT for the three-hit vs two-hit test as a function of the simulated rate ψ > 0.05` ","pinned":false,"hidden":false,"mode":"ojs"},{"id":1165,"value":"plot_power_curve (_.filter (sim_data, (d)=> d.psi > 0.05 && d.delta == 0 && d.LRT32 >= 0), [\"psi\"], 1, 0.05,\"LRT32\", [\"Three-hit rate, ψ\", \"LRT statistic (3-hit vs 2-hit)\"],\"linear\",\"median\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":1180,"value":"md`**Figure 4.** Mean (Q1-Q3) p-value for the three-hit-island vs two-hit test as a function of the simulated rate where only synonymous islands where present ψ-syn > 0.0` ","pinned":false,"hidden":false,"mode":"ojs"},{"id":1184,"value":"plot_power_curve (_.filter (sim_data, (d)=> d.delta == 0 && d.psi_syn > 0.0 && d.psi == 0), [\"psi_syn\"], 0.5, 0.05,\"i32\", [\"Three-hit rate for synonymous islands, ψ-syn\", \"p-value (3-hit-syn vs 2-hit)\"],\"linear\", \"mean\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":1243,"value":"md`**Figure 5.** Simulated rates`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1238,"value":"$2a","pinned":false,"hidden":false,"mode":"ojs"},{"id":693,"value":"$2b","pinned":false,"hidden":false,"mode":"ojs"},{"id":15,"value":"md`------`","pinned":false,"hidden":false,"mode":"ojs"},{"id":287,"value":"sim_data = {\n let s = await d3.csv (\"https://dl.dropbox.com/s/7oqyeo2cfsgrw2k/results-empirical.csv\", d3.autoType);\n _.each (s, (r)=> {\n r.null = r.delta == 0 && r.psi == 0 && r.psi_syn == 0;\n });\n return s;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":520,"value":"$2c","pinned":false,"hidden":false,"mode":"ojs"},{"id":300,"value":"$2d","pinned":false,"hidden":false,"mode":"ojs"},{"id":330,"value":"$2e","pinned":false,"hidden":false,"mode":"ojs"},{"id":381,"value":"$2f","pinned":false,"hidden":false,"mode":"ojs"},{"id":6,"value":"_=require ('lodash')","pinned":false,"hidden":false,"mode":"ojs"},{"id":9,"value":"d3 = require(\"d3\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":10,"value":"vegalite = require(\"vega-embed@3\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":514,"value":"import { table } from \"@gampleman/table\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":1217,"value":"_.filter (sim_data, (d)=> d.delta == 0 && d.psi == 0.0 && d.psi_syn > 0.0)","pinned":false,"hidden":false,"mode":"ojs"}],"files":"$W30","stdlib":"1","titling":"auto"}