-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathgeneral.html
159 lines (147 loc) · 10.3 KB
/
general.html
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
<!DOCTYPE html>
<html lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<title>Chapter 2 General | WG2: Cell-type classification</title>
<meta name="author" content="Jose Alquicira Hernandez and Lieke Michelsen">
<meta name="description" content="The cell-type annotation framework implemented in this project consists on supervised classification using two approaches: Azimuth: This method implemented in Seurat V4 classifies by cells by...">
<meta name="generator" content="bookdown 0.24 with bs4_book()">
<meta property="og:title" content="Chapter 2 General | WG2: Cell-type classification">
<meta property="og:type" content="book">
<meta property="og:url" content="https://github.com/powellgenomicslab/WG2-pipeline-classification-docs/general.html">
<meta property="og:description" content="The cell-type annotation framework implemented in this project consists on supervised classification using two approaches: Azimuth: This method implemented in Seurat V4 classifies by cells by...">
<meta name="twitter:card" content="summary">
<meta name="twitter:title" content="Chapter 2 General | WG2: Cell-type classification">
<meta name="twitter:description" content="The cell-type annotation framework implemented in this project consists on supervised classification using two approaches: Azimuth: This method implemented in Seurat V4 classifies by cells by...">
<!-- JS --><script src="https://cdnjs.cloudflare.com/ajax/libs/clipboard.js/2.0.6/clipboard.min.js" integrity="sha256-inc5kl9MA1hkeYUt+EC3BhlIgyp/2jDIyBLS6k3UxPI=" crossorigin="anonymous"></script><script src="https://cdnjs.cloudflare.com/ajax/libs/fuse.js/6.4.6/fuse.js" integrity="sha512-zv6Ywkjyktsohkbp9bb45V6tEMoWhzFzXis+LrMehmJZZSys19Yxf1dopHx7WzIKxr5tK2dVcYmaCk2uqdjF4A==" crossorigin="anonymous"></script><script src="https://kit.fontawesome.com/6ecbd6c532.js" crossorigin="anonymous"></script><script src="libs/header-attrs-2.11/header-attrs.js"></script><script src="libs/jquery-3.6.0/jquery-3.6.0.min.js"></script><meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<link href="libs/bootstrap-4.6.0/bootstrap.min.css" rel="stylesheet">
<script src="libs/bootstrap-4.6.0/bootstrap.bundle.min.js"></script><script src="libs/bs3compat-0.3.1/transition.js"></script><script src="libs/bs3compat-0.3.1/tabs.js"></script><script src="libs/bs3compat-0.3.1/bs3compat.js"></script><link href="libs/bs4_book-1.0.0/bs4_book.css" rel="stylesheet">
<script src="libs/bs4_book-1.0.0/bs4_book.js"></script><script src="https://cdnjs.cloudflare.com/ajax/libs/autocomplete.js/0.38.0/autocomplete.jquery.min.js" integrity="sha512-GU9ayf+66Xx2TmpxqJpliWbT5PiGYxpaG8rfnBEk1LL8l1KGkRShhngwdXK1UgqhAzWpZHSiYPc09/NwDQIGyg==" crossorigin="anonymous"></script><script src="https://cdnjs.cloudflare.com/ajax/libs/mark.js/8.11.1/mark.min.js" integrity="sha512-5CYOlHXGh6QpOFA/TeTylKLWfB3ftPsde7AnmhuitiTX4K5SqCLBeKro6sPS8ilsz1Q4NRx3v8Ko2IBiszzdww==" crossorigin="anonymous"></script><!-- CSS --><link rel="stylesheet" href="bs4_style.css">
</head>
<body data-spy="scroll" data-target="#toc">
<div class="container-fluid">
<div class="row">
<header class="col-sm-12 col-lg-3 sidebar sidebar-book"><a class="sr-only sr-only-focusable" href="#content">Skip to main content</a>
<div class="d-flex align-items-start justify-content-between">
<h1>
<a href="index.html" title="">WG2: Cell-type classification</a>
</h1>
<button class="btn btn-outline-primary d-lg-none ml-2 mt-1" type="button" data-toggle="collapse" data-target="#main-nav" aria-expanded="true" aria-controls="main-nav"><i class="fas fa-bars"></i><span class="sr-only">Show table of contents</span></button>
</div>
<div id="main-nav" class="collapse-lg">
<form role="search">
<input id="search" class="form-control" type="search" placeholder="Search" aria-label="Search">
</form>
<nav aria-label="Table of contents"><h2>Table of contents</h2>
<ul class="book-toc list-unstyled">
<li><a class="" href="index.html"><span class="header-section-number">1</span> About</a></li>
<li><a class="active" href="general.html"><span class="header-section-number">2</span> General</a></li>
<li><a class="" href="getting-started.html"><span class="header-section-number">3</span> Getting started</a></li>
<li><a class="" href="split-data.html"><span class="header-section-number">4</span> Split data</a></li>
<li><a class="" href="azimuth-classification.html"><span class="header-section-number">5</span> Azimuth classification</a></li>
<li><a class="" href="hierarchical-scpred-classification.html"><span class="header-section-number">6</span> Hierarchical scPred classification</a></li>
<li><a class="" href="merge-data.html"><span class="header-section-number">7</span> Merge data</a></li>
<li><a class="" href="compare-classifications.html"><span class="header-section-number">8</span> Compare classifications</a></li>
</ul>
<div class="book-extra">
<p><a id="book-repo" href="https://github.com/powellgenomicslab/WG2-pipeline-classification-docs">View book source <i class="fab fa-github"></i></a></p>
</div>
</nav>
</div>
</header><main class="col-sm-12 col-md-9 col-lg-7" id="content"><div id="general" class="section level1" number="2">
<h1>
<span class="header-section-number">2</span> General<a class="anchor" aria-label="anchor" href="#general"><i class="fas fa-link"></i></a>
</h1>
<p>The cell-type annotation framework implemented in this project consists on supervised classification using two approaches:</p>
<ul>
<li><p><strong>Azimuth</strong>: This method implemented in Seurat V4 classifies by cells by mapping the query dataset onto a supervised PCA space constructed from modal-weighted neighbors calculated by combining RNA and protein data.</p></li>
<li>
<p><strong>Hierarchical scPred:</strong> This approach is based on two methods:</p>
<ul>
<li><p><strong><em>scPred</em>:</strong> a supervised classification method that learns a low dimensional representation of a reference dataset. Unlike azimuth (which relies on a lazy classifier), <em>scPred</em> trains binary classifiers for each cell type using a <em>one vs. all</em> approach. The query dataset is projected onto the training low dimensional space and labels are assigned according to probability scores. For more details see <a href="https://genomebiology.biomedcentral.com/articles/10.1186/s13059-019-1862-5" class="uri">https://genomebiology.biomedcentral.com/articles/10.1186/s13059-019-1862-5</a></p></li>
<li><p><strong>Hierarchical progressive learning (scHPL)</strong>: this method decomposes the classification problem into smaller sub-tasks according to a hierarchy representing cell-type relationships. Cells are classified progressively using a top-bottom approach instead of using a flat classifier. For more details see <a href="https://www.nature.com/articles/s41467-021-23196-8" class="uri">https://www.nature.com/articles/s41467-021-23196-8</a></p></li>
</ul>
</li>
</ul>
<p>The cell-type classification is guided using a <a href="https://www.cell.com/cell/fulltext/S0092-8674(21)00583-3">CITE-seq reference</a> containing:</p>
<ul>
<li>161,764 human PBMCs</li>
<li>20,953 genes</li>
<li>224 protein markers (CITE-seq)</li>
<li>31 cell types</li>
</ul>
<div class="inline-table"><table class="table table-sm">
<colgroup>
<col width="71%">
<col width="28%">
</colgroup>
<thead><tr class="header">
<th><strong>Lymphoid populations</strong></th>
<th><strong>Myeloid populations</strong></th>
</tr></thead>
<tbody>
<tr class="odd">
<td>CD4 Naive, CD4 TCM, CD4 TEM, CD4 CTL, Treg, CD4 Proliferating</td>
<td>CD14 Mono, CD16 Mono</td>
</tr>
<tr class="even">
<td>CD8 Naive, CD8 TCM, CD8 TEM, CD8 Proliferating</td>
<td>cDC1, cDC2, pDC, ASDC</td>
</tr>
<tr class="odd">
<td>gdT, MAIT, ILC, dnT</td>
<td>HSPC</td>
</tr>
<tr class="even">
<td>NK, NK_CD56bright, NK Proliferating</td>
<td>Platelet, Eryth</td>
</tr>
<tr class="odd">
<td>B naive, B intermediate, B memory, Plasmablast</td>
<td>Doublet</td>
</tr>
</tbody>
</table></div>
<blockquote>
<p>“This PBMC reference dataset was generated as part of the Hao and Hao
et al, Cell 2021 paper. It is comprised of data from eight volunteers
enrolled in an HIV vaccine trial from which three time point samples
were taken at day 0, 3, and 7 following vaccination. All 24 samples
were processed with a CITE-seq panel of 228 TotalSeq A antibodies to
generate single-cell RNA and ADT data. The data were then integrated
using metholody described in the pre-print linked above to generate a
weighted nearest neighbor (WNN) representation of the RNA and protein
data jointly. This WNN representation is used in the Azimuth app to
assign celltypes, embed in the reference UMAP, and impute protein
levels for the query dataset”. <em>Hao Y, et al. Cell. (2021)</em></p>
</blockquote>
</div>
<div class="chapter-nav">
<div class="prev"><a href="index.html"><span class="header-section-number">1</span> About</a></div>
<div class="next"><a href="getting-started.html"><span class="header-section-number">3</span> Getting started</a></div>
</div></main><div class="col-md-3 col-lg-2 d-none d-md-block sidebar sidebar-chapter">
<nav id="toc" data-toggle="toc" aria-label="On this page"><h2>On this page</h2>
<ul class="nav navbar-nav"><li><a class="nav-link" href="#general"><span class="header-section-number">2</span> General</a></li></ul>
<div class="book-extra">
<ul class="list-unstyled">
<li><a id="book-source" href="https://github.com/powellgenomicslab/WG2-pipeline-classification-docs/blob/master/02-general.Rmd">View source <i class="fab fa-github"></i></a></li>
<li><a id="book-edit" href="https://github.com/powellgenomicslab/WG2-pipeline-classification-docs/edit/master/02-general.Rmd">Edit this page <i class="fab fa-github"></i></a></li>
</ul>
</div>
</nav>
</div>
</div>
</div> <!-- .container -->
<footer class="bg-primary text-light mt-5"><div class="container"><div class="row">
<div class="col-12 col-md-6 mt-3">
<p>"<strong>WG2: Cell-type classification</strong>" was written by Jose Alquicira Hernandez and Lieke Michelsen. It was last built on 2022-06-14.</p>
</div>
<div class="col-12 col-md-6 mt-3">
<p>This book was built by the <a class="text-light" href="https://bookdown.org">bookdown</a> R package.</p>
</div>
</div></div>
</footer>
</body>
</html>