generated from jhudsl/OTTR_Template
-
Notifications
You must be signed in to change notification settings - Fork 6
/
Copy path07-microarray-data.Rmd
191 lines (116 loc) · 12.4 KB
/
07-microarray-data.Rmd
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
```{r, include = FALSE}
ottrpal::set_knitr_image_path()
```
# Microarray Data
<div class = "warning">
This chapter is in a beta stage. If you wish to contribute, please [go to this form](https://forms.gle/dqYgmKH8XXE2ohwD9) or our [GitHub page](https://github.com/fhdsl/Choosing_Genomics_Tools).
</div>
## Learning Objectives
```{r, fig.alt = "This chapter will demonstrate how to: Understand the very general basics of microarray data collection and processing workflow. Understand the limitations and strengths of microarray data in general.", out.width = "100%", echo = FALSE}
ottrpal::include_slide("https://docs.google.com/presentation/d/1YwxXy2rnUgbx_7B7ENH9wpDX-j6JpJz6lGVzOkjo0qY/edit#slide=id.g144bc8d6a68_0_12")
```
## Summary of microarrays
Microarrays have been in use since before high throughput sequencing methods became more affordable and widespread, but they still can be a effective and affordable tool for genomic assays. Depending on your goals, microarray may be a suitable choice for your genomic study.
## How do microarrays work?
All microarrays work on hybridization to sets of oligonucleotides on a chip. However, the preparation of the samples, and the oligonucleotides' hybridization targets vary depending on the assay and goals.
```{r, fig.alt = "", out.width = "100%", echo = FALSE}
ottrpal::include_slide("https://docs.google.com/presentation/d/1YwxXy2rnUgbx_7B7ENH9wpDX-j6JpJz6lGVzOkjo0qY/edit#slide=id.g14492c87338_0_51")
```
On a basic principle, oligonucleotide probes are designed for different targets sets designed for the same targets are put together. On the whole chip, these probes are arranged in a grid like design so that after a sample is hybridized to them, you can detect how much of the target is detected by taking an image and knowing what target each location is designed to.
### Pros:
- Microarrays are much more affordable than high throughput sequencing which can allow you to run more samples and have more statistical power [@Tarca2006; @refinebioexamples2019].
- Microarrays take less time to process than most high throughput sequencing methods[@Tarca2006; @refinebioexamples2019].
- Microarrays are generally less computationally intensive to process and you can get your results more quickly[@Tarca2006; @refinebioexamples2019].
- Microarrays are generally as good as sequencing methods for detecting clinical endpoints [@Zhang2015].
### Cons:
- Microarray chips can only measure the targets they are designed for, and cannot be used for exploratory purposes [@Zhang2015].
- Microarrays' probe designs can only be as up to date as the genome they were designed against at the time [@Mantione2014; @refinebioexamples2019].
- Microarray does not escape oligonucleotide biases like GC content and sequence composition biases[@refinebioexamples2019].
## What types of arrays are there?
### SNP arrays
```{r, fig.alt = "", out.width = "100%", echo = FALSE}
ottrpal::include_slide("https://docs.google.com/presentation/d/1YwxXy2rnUgbx_7B7ENH9wpDX-j6JpJz6lGVzOkjo0qY/edit#slide=id.g13438a9a5b2_0_16")
```
Single nucleotide polymorphism arrays are designed to measure DNA variants. They are designed to target DNA variants. When the sample is hybridized, the amount of fluorescence detected can be interpreted to indicate the presence of the variant and whether the variant is homogeneous or heterogenous. The samples prepped for SNP arrays then need to be DNA samples.
#### Examples:
- [The 1000 genomes project](https://www.internationalgenome.org/) is a large collection of SNP data array from many populations around the world and is available for download.
### Gene expression arrays
```{r, fig.alt = "", out.width = "100%", echo = FALSE}
ottrpal::include_slide("https://docs.google.com/presentation/d/1YwxXy2rnUgbx_7B7ENH9wpDX-j6JpJz6lGVzOkjo0qY/edit#slide=id.g142e3de7ce8_0_8")
```
Gene expression arrays are designed to measure gene expression. They are designed to target and measure relative transcript abundance level.
#### Examples:
- [refine.bio](https://www.refine.bio/) is the largest collection of publicly available, already normalized gene expression data (including gene expression microarrays).
- [Getting started in gene expression microarray analysis](https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1000543) [@Slonim_Yanai_2009].
- [Microarray and its applications](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3467903/) [-@Govindarajan2012].
- [Analysis of microarray experiments of gene expression profiling](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2435252/) [@Tarca2006].
### DNA methylation arrays
```{r, fig.alt = "", out.width = "100%", echo = FALSE}
ottrpal::include_slide("https://docs.google.com/presentation/d/1YwxXy2rnUgbx_7B7ENH9wpDX-j6JpJz6lGVzOkjo0qY/edit#slide=id.g144bc8d6a68_0_1")
```
DNA methylation can also be measured by microarray. To detect methylated cytosines (5mC), DNA samples are prepped using bisulfite conversion. This converts unmethylated cytosines into uracils and leaves methylated cytosines untouched. Probes are then designed to bind to either the uracil or the cytosine, representing the unmethylated and methylated cytosines respectively.
A ratio of the fluorescence signal can be used to identify the relative abundance of the methylated and unmethylated versions of the sequence.
Additionally, 5-hydroxymethylated cytosines (5hmC) can also be detected by oxidative bisulfite bisulfite sequencing [@Booth2013]. Note that bisulfite conversion alone will not distinguish between 5mC and 5hmC though these often may indicate different biological mechanics.
## General processing of microarray data
```{r, fig.alt = "", out.width = "100%", echo = FALSE}
ottrpal::include_slide("https://docs.google.com/presentation/d/1YwxXy2rnUgbx_7B7ENH9wpDX-j6JpJz6lGVzOkjo0qY/edit#slide=id.g15f36710259_0_0")
```
After scanning, microarray data starts as an image that needs to be quantified, normalized and further corrected and edited based on the most current genome and probe annotation.
As noted above, microarrays do not escape the base sequence biases that accompany most all genomic assays. The normalization methods you use ideally will mitigate these sequence biases and also make sure to remove probes that may be outdated or bind to multiple places on the genome.
The tools and methods by which you normalize and correct the microarray data will be dependent not only on the type of microarray assay you are performing (gene expression, SNP, methylation), but most of all what kind of microarray chip design/platform you are using.
### Examples
- [Refine.bio describes their processing methods](https://docs.refine.bio/en/latest/main_text.html#processing-information).
- [Brainarray keeps up to date microarray annotation for all kinds of platforms](http://brainarray.mbni.med.umich.edu/Brainarray/Database/CustomCDF/genomic_curated_CDF.asp)
### Microarray Platforms
There are so many microarray chip designs out there designed to target different things. Three of the largest commercial manufacturers have ready to use microarrays you can purchase. You can also design microarrays to hit your own targets of interest.
Here are full lists of platforms that have been published on Gene Expression Omnibus.
- [Affymetrix platforms](https://www.ncbi.nlm.nih.gov/geo/browse/?view=platforms&submitter=5&zsort=series&display=20)
- [Agilent platforms](https://www.ncbi.nlm.nih.gov/geo/browse/?view=platforms&submitter=63&display=20&zsort=series).
- [Illumina platforms](https://www.ncbi.nlm.nih.gov/geo/browse/?view=platforms&submitter=1242&zsort=series&display=20).
## Very General Microarray Workflow
In the data type specific chapters, we will cover the microarray workflow and file formats in more detail. But in the most general sense, microarray workflows look like this, note that the exact file formats are specific to the chip brand and type you use (e.g. Illumina, Affymetrix, Agilent, etc.):
<div id="CB1322FC70C4A4B3A1BF2D8AC29A7E9FFF8_81516"><div id="CB1322FC70C4A4B3A1BF2D8AC29A7E9FFF8_81516_robot"><a href="https://cloud.smartdraw.com/share.aspx/?pubDocShare=CB1322FC70C4A4B3A1BF2D8AC29A7E9FFF8" target="_blank"><img src="https://cloud.smartdraw.com/cloudstorage/CB1322FC70C4A4B3A1BF2D8AC29A7E9FFF8/preview2.png"></a></div></div><script src="https://cloud.smartdraw.com/plugins/html/js/sdjswidget_html.js" type="text/javascript"></script><script type="text/javascript">SDJS_Widget("CB1322FC70C4A4B3A1BF2D8AC29A7E9FFF8",81516,1,"");</script><br/>
### Microarray file formats
#### IDAT - intensity data file
This is an Illumina microarray specific file that contains the chip image intensity information for each location on the microarray. It is a binary file, which means it will not be readable by double clicking and attempting to open the file directly.
Currently, Illumina appears to suggest directly converting IDAT files into a GTC format. We advise looking into [this package to help you do that](https://github.com/freeseek/gtc2vcf).
[For more on IDAT files](https://www.illumina.com/content/dam/illumina-marketing/documents/products/technotes/technote_array_analysis_workflows.pdf).
#### DAT - data file
This is an Affymetrix' microarray specific file parallel to the IDAT file in that it contains the image intensity information for each location on the microarray. It's stored as pixels.
[For more on DAT files](https://www.affymetrix.com/support/developer/powertools/changelog/gcos-agcc/dat.html).
#### CEL
This is an Affymetrix microarray specific file that is made from a DAT file but translated into numeric values. It is not normalized yet but can be normalized into a CHP file.
[For more on CEL files](https://www.affymetrix.com/support/developer/powertools/changelog/gcos-agcc/cel.html)
#### CHP
CHP files contain the gene-level and normalized data from an Affymetrix array chip. CHP files are obtained by normalizing and processing CEL files.
[For more about CHP files](https://www.affymetrix.com/support/developer/powertools/changelog/gcos-agcc/chp-xda.html).
## General informatics files
At various points in your genomics workflows, you may need to use other types of files to help you annotate your data. We'll also discuss some of these common files that you may encounter:
#### BED - Browser Extensible Data
A BED file is a text file that has coordinates to genomic regions. THe other columns that accompany the genomic coordinates are variable depending on the context. But every BED file contains the `chrom`, `chromStart` and `chromEnd` columns to start.
A BED file might look like this:
```
chrom chromStart chromEnd other_optional_columns
chr1 0 1000 good
chr2 100 3000 bad
```
For [more on BED files](https://en.wikipedia.org/wiki/BED_(file_format)).
#### GFF/GTF General Feature Format/Gene Transfer Format
A GFF file is a tab delimited file that contains information about genomic features. These types of files are available from databases and what you can use to annotate your data.
You may see there are GFF2, GFF3, and GTF files. These only refer to different versions and variations. They generally have the same information. In general, GFF2 is being phased out so using GFF3 is generally a better bet unless the program or package you are using specifies it needs an older GFF2 version.
A GFF file may look like this (borrowed example from Ensembl):
```
1 transcribed_unprocessed_pseudogene gene 11869 14409 . + . gene_id "ENSG00000223972"; gene_name "DDX11L1"; gene_source "havana"; gene_biotype "transcribed_unprocessed_pseudogene";
```
Note that it will be useful for annotating genes and what we know about them.
For [more about GTF and GFF files](https://useast.ensembl.org/info/website/upload/gff.html).
### Other files
\* If you didn't see a file type listed you are looking for, take a look at this [list by the BROAD](https://software.broadinstitute.org/cancer/software/gsea/wiki/index.php/Data_formats). Or, it may be covered in the data type specific chapters.
### Microarray processing tutorials:
For the most common microarray platforms, you can see these examples for how to process the data:
#### General arrays
- [Using Bioconductor for Microarray Analysis](https://www.bioconductor.org/packages/devel/workflows/vignettes/arrays/inst/doc/arrays.html).
#### Gene Expression Arrays
- [An end to end workflow for differential gene expression using Affymetrix microarrays](https://f1000research.com/articles/5-1384).
#### DNA Methylation Arrays
- [DNA Methylation array workflow](https://nbis-workshop-epigenomics.readthedocs.io/en/latest/content/tutorials/methylationArray/Array_Tutorial.html).