31:T443,//ggplot(filtered_data, aes(x=bin, y=MEAN_PROBABILITY)) + geom_boxplot(fill='lightblue') + coord_flip() + xlab("Branch Lengths (MG94XREV)") vegalite({ "$schema": "https://vega.github.io/schema/vega-lite/v5.json", "description": "A vertical box plot showing median, min, and max mean probability scores.", "data": {"values": filtered_results}, "width" : "1000", "height" : "300", "mark": { "type": "boxplot", "extent": "min-max" }, "encoding": { "x": {"field": "bin", "type": "nominal", 'sort': ["[0,0.01]","(0.01,0.02]","(0.02,0.03]","(0.03,0.04]","(0.04,0.05]","(0.05,0.06]","(0.06,0.07]","(0.07,0.08]","(0.08,0.09]","(0.09,0.1]","(0.1,0.11]","(0.11,0.12]","(0.12,0.13]","(0.13,0.14]","(0.14,0.15]","(0.15,0.16]","(0.16,0.17]","(0.17,0.18]","(0.18,0.19]","(0.19,0.2]"], "title":"Branch Lengths (MG94XREV)"}, //"color": {"field": "bin", "type": "nominal", "legend": null}, "y": { "field": "MEAN_PROBABILITY", "title": "Mean Probability", "type": "quantitative", "scale": {"zero": false} } } } )32:T409,vegalite({ "$schema": "https://vega.github.io/schema/vega-lite/v5.json", "data": {"values": avg_missing_codons}, // "transform": [ // {"calculate": "{'CAT': 'A', 'HORSE': 'B', 'RAT': 'C', 'BABOON': 'D', 'COW': 'E', 'RHMONKEY':'F', 'CHIMP':'G', 'HUMAN': 'H' }[datum.animal]", "as": "animal"} // ], "width" : "1000", "height" : "500", "mark": "bar", "encoding": { "x": {"field": "bin", 'sort': ["[0,0.01]","(0.01,0.02]","(0.02,0.03]","(0.03,0.04]","(0.04,0.05]","(0.05,0.06]","(0.06,0.07]","(0.07,0.08]","(0.08,0.09]","(0.09,0.1]","(0.1,0.11]","(0.11,0.12]","(0.12,0.13]","(0.13,0.14]","(0.14,0.15]","(0.15,0.16]","(0.16,0.17]","(0.17,0.18]","(0.18,0.19]","(0.19,0.2]"], axis: { title: "Branch Lengths (MG94XREV)" }}, "y": {"aggregate": "sum", "field": "avg_missing", axis: { title: "Average # of times true codon was not in prediction set" }}, //"scale": {"domain": [0, 50]}, // "color": {"field": "animal"}, "tooltip": {"field": "avg_missing", "type": "quantitative"} } })33:[{"id":"057d29e6e9d01aff6ce5fde05149408cfc79b7d6c2e29e7c87f8f280e28c97b7f195647cdf0d578cb292421d1547fdbbc30a79a1a21a1d6e3d74348c1b78d999","href":"https://static.observableusercontent.com/files/057d29e6e9d01aff6ce5fde05149408cfc79b7d6c2e29e7c87f8f280e28c97b7f195647cdf0d578cb292421d1547fdbbc30a79a1a21a1d6e3d74348c1b78d999","name":"empirical-trees@5.json","type":"application/json","lastModified":1652134044908.107,"size":26666},{"id":"a576d60205e9ea14450ce4f575ce43c1197f1c8ff1634fa7108a6dd552bb7c91dd440118e8e8dc0686378e5db9bd36edd534fdc1876c7e830519c553f119e932","href":"https://static.observableusercontent.com/files/a576d60205e9ea14450ce4f575ce43c1197f1c8ff1634fa7108a6dd552bb7c91dd440118e8e8dc0686378e5db9bd36edd534fdc1876c7e830519c553f119e932","name":"empirical-results-20220413.json","type":"application/json","lastModified":1651610224978.509,"size":221198}] 2f:{"title":"PREI Evolutionary imputation validation:","cells":[{"id":0,"value":"# PREI Evolutionary imputation validation:\n## empirical dataset","pinned":false,"hidden":false,"mode":"md"},{"id":795,"value":"Takeaway: PRIME performance on empirical data is similar to performance on simulated data.","pinned":false,"hidden":false,"mode":"md"},{"id":716,"value":"## **Figure 1.** Histogram of branch lengths \nMost alignments have shorter branches (a roughly equal number of short and medium-length branches). Binwidth = 0.01 ","pinned":false,"hidden":false,"mode":"md"},{"id":655,"value":"Plot.plot({\n x: {label: 'Branch Length'},\n y: {\n grid: true\n },\n color: {\n legend: true\n },\n marks: [\n Plot.rectY(full_results, Plot.binX({y: \"count\"}, {x: {interval: 0.01, value: 'MG94XREV'}, fill: \"alignment\"})),\n //{x: 'MG94XREV', fill: \"alignment\"} doesn't work\n Plot.ruleY([0])\n ],\n width: 1000\n})\n\n// chart = Histogram(full_results, {\n// value: d => d.MG94XREV,\n// label: \"Branch Length →\",\n// width,\n// height: 500,\n// fill: full_results.alignment\n// })","pinned":false,"hidden":false,"mode":"ojs"},{"id":742,"value":"## **Figure 2.** Predictive ability decreases as branch length increases \nBranch lengths > 0.2 were filtered out, as there was not enough data in each bin for the boxplot to be informative. ","pinned":false,"hidden":false,"mode":"md"},{"id":746,"value":"$31","pinned":false,"hidden":false,"mode":"ojs"},{"id":800,"value":"## **Figure 3.** Average number of times the true codon wasn't in the set of PRIME predictions for different branch lengths.","pinned":false,"hidden":false,"mode":"md"},{"id":990,"value":"$32","pinned":false,"hidden":false,"mode":"ojs"},{"id":250,"value":"\nviewof chosen_dataset = select ({\n \"options\" : [\"adh.nex\", \"bglobin.nex\", \"camelid.nex\", \"ENCenv.nex\" , \"flavNS5.nex\", \"HepatitisD.nex\", \"HIVvif.nex\", \"IAV_Chen.fasta\", \"lysin.nex\", \"COXI.mtnex\", \"yokoyama.rh1.cds.mod.1-990.nex\"],\n \"title\" : \"Choose dataset:\"\n});","pinned":false,"hidden":false,"mode":"ojs"},{"id":283,"value":"md`### Tree for ${chosen_dataset}\nNote: Branch lengths were estimated in hyphy with Global MG94xREV.`","pinned":false,"hidden":false,"mode":"ojs"},{"id":400,"value":"chosen_tree = { \n let chosen_tree = \"\"\n for (let i = 0; i < data_and_trees.length; i++) {\n if (data_and_trees[i].DATASET == chosen_dataset){ \n return chosen_tree = data_and_trees[i].TREE;\n}\n}\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":329,"value":"rendered_tree = tree.render({\n height:chosen_data.length*7, \n width:700,\n 'left-right-spacing': 'fit-to-size', \n 'top-bottom-spacing': 'fit-to-size'\n});","pinned":false,"hidden":false,"mode":"ojs"},{"id":501,"value":"md`#### How branch length affects mean probability for the dataset: ${chosen_dataset}`","pinned":false,"hidden":false,"mode":"ojs"},{"id":436,"value":"Plot.plot({\n y: {\n grid: true,\n inset: 5\n },\n marks: [\n Plot.lineY(chosen_data, \n {x: \"MG94XREV\", \n y: \"MEAN_PROBABILITY\", \n stroke: \"property\",\n sort: 'MG94XREV'}),\n Plot.dot(chosen_data, {x: \"MG94XREV\", y: \"MEAN_PROBABILITY\", fill: \"property\"})\n ],\n color: {\n legend: true\n },\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":516,"value":"md`#### How property affects mean probability for the dataset: ${chosen_dataset}`","pinned":false,"hidden":false,"mode":"ojs"},{"id":469,"value":"Plot.plot({\n marks: [\n Plot.boxY(chosen_data, {x: \"property\", y: \"MEAN_PROBABILITY\", fill: \"property\"}),\n Plot.dot(chosen_data, {x: \"property\", y: \"MEAN_PROBABILITY\"})\n ],\n color: {\n legend: true\n }\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":1106,"value":"#### Relationship between the PRIME algorithm's predictive ability and branch length for the dataset: ${chosen_dataset}\nPRIME's prediction set nearly always contains the true codon. Branches for which PRIME does not predict the true codon at one or more sites tend to have a wide distribution of lengths in many datasets. The lack of correlation between branch length and PRIME's prediction accuracy indicates that branch length is not a major factor in PRIME's failure to predict the true codon.","pinned":false,"hidden":false,"mode":"md"},{"id":526,"value":"Plot.plot({\n x: {label: \"Number of times the true codon was not in set of PRIME predictions\"},\n y: {\n grid: true,\n inset: 5,\n label: 'Branch Length'\n },\n marks: [\n Plot.boxY(chosen_data_missing_codons, {x: \"NOT_PREDICTED\", y: \"MG94XREV\", fill: 'orange'}),\n Plot.dot(chosen_data_missing_codons, {y: \"MG94XREV\", x: \"NOT_PREDICTED\"}), \n ],\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":753,"value":"**Import Datasets**","pinned":false,"hidden":false,"mode":"md"},{"id":338,"value":"data_and_trees = FileAttachment(\"empirical-trees@5.json\").json()","pinned":false,"hidden":false,"mode":"ojs"},{"id":871,"value":"full_results_json = FileAttachment(\"empirical-results-20220413.json\").json()","pinned":false,"hidden":false,"mode":"ojs"},{"id":816,"value":"full_results = await aq.loadCSV('https://raw.githubusercontent.com/hverdonk/evolutionary-imputation-validation/master/empirical-results-20220413.csv')","pinned":false,"hidden":false,"mode":"ojs"},{"id":631,"value":"//chosen_data = full_results.ungroup().filter(d => d.alignment == chosen_dataset).orderby('MG94XREV').reify()\n\nchosen_data = { \nconst reduced = [];\nfor (let i = 0; i < full_results_json.length; i++) {\nif (full_results_json[i].alignment == chosen_dataset){ \n reduced.push(full_results_json[i]);\n}\n}\n return reduced\n}\n\n","pinned":false,"hidden":false,"mode":"ojs"},{"id":1091,"value":"chosen_data_missing_codons = { \nconst reduced = [];\nfor (let i = 0; i < chosen_data.length; i++) {\n let new_row = {...chosen_data[i], ...{NOT_PREDICTED: chosen_data[i].N - chosen_data[i].IN_SET}};\n reduced.push(new_row);\n// if (full_results_json[i].alignment == chosen_dataset){ \n// reduced.push(full_results_json[i]);\n// }\n}\n return reduced\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":411,"value":"// Filtered out branch lengths > 0.2 (not enough branches in each bin for the global boxplot to be informative). \n// filtered_results = FileAttachment(\"filtered_BLlessequal0.2.json\").json()\nfiltered_results = await aq.loadCSV('https://raw.githubusercontent.com/hverdonk/evolutionary-imputation-validation/master/empirical-results-20220413-filtered-BLlessequal0.2.csv')","pinned":false,"hidden":false,"mode":"ojs"},{"id":805,"value":"// we can do all sorts of dplyr-esq things\n// Tip: rollup is like summarize; derive is like mutate\navg_missing_codons = filtered_results\n .groupby('bin')\n .derive({\n missing: (d => d.N-d.IN_SET),\n })\n .rollup({\n count: op.count(),\n total_missing: op.sum('missing')\n })\n .derive({avg_missing: (d => d.total_missing/d.count),})","pinned":false,"hidden":false,"mode":"ojs"},{"id":320,"value":"tree = new phylotree.phylotree(chosen_tree);","pinned":false,"hidden":false,"mode":"ojs"},{"id":756,"value":"**Import Code**","pinned":false,"hidden":false,"mode":"md"},{"id":47,"value":"import {applyHljsStyle} from '@anwarhahjjeffersongeorge/code-block-style-picker'","pinned":false,"hidden":false,"mode":"ojs"},{"id":281,"value":"import {select, checkbox, autoSelect} from \"@jashkenas/inputs\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":656,"value":"import {Histogram} from \"@d3/histogram\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":313,"value":"d3 = require(\"d3@5\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":750,"value":"vegalite = require (\"vega-embed\")","pinned":true,"hidden":false,"mode":"ojs"},{"id":808,"value":"import {aq, op} from '@uwdata/arquero'","pinned":false,"hidden":false,"mode":"ojs"},{"id":316,"value":"phylotree=require ('phylotree@1.0.0-alpha.25')","pinned":false,"hidden":false,"mode":"ojs"},{"id":331,"value":"stylesheet = html``","pinned":false,"hidden":false,"mode":"ojs"}],"files":"$W33","stdlib":"1","titling":"auto"}