Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 | 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 16x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 284x 29x 255x 2x 2x 2x 2x | /**
* A track's values compared between groups of the genome, as a descriptor.
*
* The question behind every *is this compartment GC-rich*, *is B transcribed
* less*, *is the signal higher at these peaks than at those*: one quantity, read
* over sets of places, and the sets compared. Every piece of it was already
* here — a compartment call leaves its A and B as intervals, any lane is a
* track — and what was missing was the comparison itself, which a reader did by
* exporting both and writing a script.
*
* ## Where the groups come from
*
* **A feature set**, split by what is written on its features — the name,
* which is A and B on a compartment call and a class on a peak set; the strand;
* the score, in quantiles; or nothing, one group — with what the set does not
* cover as a group of its own. Or **a track**, its bins cut by sign — an
* eigenvector's two compartments without calling them first — or into
* quantiles, the stratification a saddle plot makes of the same numbers. A
* grouping made from a track is written out as intervals, named by group, so the
* next task can read them as a set.
*
* ## What is compared
*
* One value per entry, by default: each run of a group's features, merged where
* they overlap or touch, each stretch of the rest between them, and each run of
* bins in one group of a track — so a 30 kb compartment is one observation, as
* a 300 bp peak is. Neighbouring bins of one compartment share their GC, their
* expression and their place in the nucleus, and a test that counted each as
* independent gave a P that followed the bin size rather than the data. A bin
* size compares stretches of at most one bin instead: a group's merged
* intervals cut on the bin grid, a stretch shorter than half a bin joined to its
* neighbour, and a feature shorter than a bin kept whole. Either way each is the
* track's mean over the bases it has a value for. A track is cut on the open
* map's resolution, and on 10 kb with no map open, unless a bin size is given.
*
* The groups are summarised, drawn as violins, and tested pairwise and together
* by ranks (`analysis/groupstats.ts`), with the q-values of every test in the
* run adjusted together.
*/
import type { AnalysisContext, AnalysisTask, CostEstimate, Params, ParamSpec } from '../types'
import { READY, waitingFor } from '../types'
/** The bins a track is cut on with no map open, in base pairs — base composition's too. */
export const GROUP_BINSIZE = 10_000
/** The bins a context gives a track: the open map's resolution, else 10 kb. */
function gridOf(context: AnalysisContext): number {
return context.hasMap && context.binsize > 0 ? context.binsize : GROUP_BINSIZE
}
const byTrack = (params: Params): boolean => params.groupBy === 'track'
const byFeatures = (params: Params): boolean => !byTrack(params)
const PARAMS: readonly ParamSpec[] = [
{
kind: 'track',
key: 'tracks',
label: 'tracks',
multiple: true,
hint: 'The values compared between the groups.',
},
{
kind: 'choice',
key: 'groupBy',
label: 'groups from',
default: 'features',
options: [
{
value: 'features',
label: 'a feature set',
hint: 'Its intervals, split by what is written on them: A and B on a compartment call.',
},
{
value: 'track',
label: 'a track',
hint: 'Its bins, cut by sign or into quantiles: an eigenvector’s two compartments.',
},
],
},
{ kind: 'featureSet', key: 'features', label: 'features', when: byFeatures },
{
kind: 'choice',
key: 'split',
label: 'split the set by',
default: 'name',
options: [
{ value: 'name', label: 'name', hint: 'One group per name the features carry.' },
{ value: 'strand', label: 'strand' },
{ value: 'score', label: 'score, in quantiles' },
{ value: 'none', label: 'nothing: one group' },
],
when: byFeatures,
},
{
kind: 'toggle',
key: 'outside',
label: 'the rest as a group',
hint: 'Every base the set does not cover, as one more group.',
default: true,
when: byFeatures,
},
{
kind: 'track',
key: 'groupTrack',
label: 'grouping track',
optional: true,
startsEmpty: true,
noneLabel: 'choose one',
when: byTrack,
},
{
kind: 'choice',
key: 'cut',
label: 'cut it by',
default: 'sign',
options: [
{ value: 'sign', label: 'sign', hint: 'Above zero and below it.' },
{ value: 'quantiles', label: 'quantiles', hint: 'Equal numbers of bins, lowest to highest.' },
],
when: byTrack,
},
{
kind: 'number',
key: 'quantiles',
label: 'quantiles',
default: 4,
min: 2,
max: 10,
integer: true,
when: (params) =>
(byTrack(params) && params.cut === 'quantiles') || (byFeatures(params) && params.split === 'score'),
},
{
kind: 'number',
key: 'resolution',
label: 'bin size',
unit: 'bp',
hint:
'0 compares one value per entry — a run of features, the stretch between two, a run of bins ' +
'in one group — and any other size compares stretches of at most that many bases.',
default: 0,
min: 0,
integer: true,
},
{
kind: 'toggle',
key: 'perFeature',
label: 'score every feature',
hint: 'A row per feature of the set, its mean on every track.',
default: false,
when: byFeatures,
},
]
function estimate(context: AnalysisContext, params: Params): CostEstimate {
const tracks = Array.isArray(params.tracks) ? params.tracks.length : 0
const free = { class: 'free' as const, requests: 0, bytes: 0, note: 'Nothing to read.' }
if (tracks === 0) return { ...free, refusal: 'Choose at least one track to compare.' }
const chosen = (key: string): boolean => typeof params[key] === 'string' && params[key] !== ''
if (byFeatures(params) && !chosen('features')) {
return { ...free, refusal: 'Choose the feature set whose groups to compare.' }
}
if (byTrack(params) && !chosen('groupTrack')) {
return { ...free, refusal: 'Choose the track to group the bins by.' }
}
const read = `${tracks.toLocaleString()} track${tracks === 1 ? '' : 's'} read`
const perEntry = !(typeof params.resolution === 'number' && params.resolution > 0)
if (perEntry && byFeatures(params)) {
return { class: 'bounded', requests: null, bytes: null, note: `${read} once over each entry of the groups` }
}
const binsize = perEntry ? gridOf(context) : (params.resolution as number)
let bases = 0
for (const contig of context.contigs) bases += contig.length
const stretches = bases > 0 ? Math.ceil(bases / binsize) : 0
return {
class: 'bounded',
requests: null,
bytes: null,
note:
`${read} over up to ${stretches.toLocaleString()} ${perEntry ? 'bins' : 'stretches'} of ` +
`${binsize.toLocaleString()} bp${perEntry ? ', compared a run of one group at a time' : ''}`,
}
}
export const GROUP_SIGNAL_TASK: AnalysisTask = {
id: 'groupsignal',
title: 'Signal by group',
blurb: 'A track’s values in each group of the genome — A against B — compared and tested.',
keywords: ['A versus B', 'compartment signal', 'box plot', 'test'],
group: '1d',
// Tracks and interval sets, which live in the annotations worker; no map is
// needed, so a 1D browser runs it as well as a map does.
venue: 'annotations',
// 2: a bin size of 0, the default, is one value per entry, where it was the
// map's resolution.
version: 2,
// A track is all it needs: a second track, or the same one, can do the
// grouping, so a feature set is one way in rather than a requirement.
availability: (context) =>
context.entries.some((entry) => entry.kind === 'track')
? READY
: waitingFor('Load a bigWig, or run a task that makes a track: the values to compare.'),
params: PARAMS,
outputs: ['curve', 'table', 'intervals'],
estimate,
}
|