-
Notifications
You must be signed in to change notification settings - Fork 0
/
slides.Rmd
446 lines (264 loc) · 11.5 KB
/
slides.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
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
% Functional Genomics
% David Montaner
% [www.dmontaner.com](http://www.dmontaner.com)
From Genes to Phenotype
==============================================================
--------------------------------------------------------------
![](images/dna2phenotype.png)
--------------------------------------------------------------
![](images/path1.png)
--------------------------------------------------------------
![](images/path2.png)
--------------------------------------------------------------
![](images/path3.png)
--------------------------------------------------------------
![](images/dna2phenotype.png)
Annotation Data Bases - Gene Ontology
--------------------------------------------------------------
![](images/dag.gif)
--------------------------------------------------------------
![](images/dna2phenotype.png)
Experimental Data: Microarrays & NGS
--------------------------------------------------------------
![](images/array.png)
thousands or millions of genomic _variables_
--------------------------------------------------------------
![](images/expression.png)
Gene Expression Analysis
==============================================================
Clustering
--------------------------------------------------------------
![](images/clustering.png)
Class prediction
--------------------------------------------------------------
![](images/predictors.jpg)
Hypothesis testing
--------------------------------------------------------------
![](images/difexp.png)
--------------------------------------------------------------
![](images/dag.gif)
Gene Set Analysis
==============================================================
--------------------------------------------------------------
![](images/over_representation_methods.png)
Shift in the observation unit
--------------------------------------------------------------
- $H_0: gene_i$ is not differentially expressed ...
- $H_0: gene\ set_j$ is not __enriched__ ...
<hr>
1. more interpretable for researchers
1. greater power (sample size )
1. smaller p-value adjustment:
20,000 genes -> 2,000 GO terms
Takes advantage of the mildly differentially expressed genes
GSEA
--------------------------------------------------------------
![](images/gsea.jpg)
Drawbacks
--------------------------------------------------------------
- Confidence level based on a __sample__ _permutation_ test.
Computationally intensive.
- Completely dependent of the gene level statistic.
Not very flexible: weighting, covariates, ...
- Tackles just a single _genomic dimension_
Genomic dimensions
--------------------------------------------------------------
- Gene expression
- Transcript expression
- Splicing
- microRNA activity
- Promoters
- Methylation
- SNV (genetic variation)
- Copy Number
- Loss of heterozygosity (LOH)
Logistic Regression Model
==============================================================
--------------------------------------------------------------
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha \ r_i $$
\
We model the probability of a gene belonging to a GO term
as a function of its _score_
in the differential expression analysis
--------------------------------------------------------------
![](images/over_representation_methods.png)
--------------------------------------------------------------
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha \ r_i $$
\
\begin{align}
&\alpha > 0 \rightarrow \text{enrichment is } + \\
&\alpha < 0 \rightarrow \text{enrichment is } - \\
&\alpha = 0 \rightarrow \text{no enrichment}
\end{align}
Ranking index
--------------------------------------------------------------
Our proposal:
$$ r = - sign (statistic) \cdot log (pvalue) $$
makes results more comparable across different studies
(good for meta analysis)
_quantiles_ transformation to a $N(0,1)$
reduces the effect of outliers
\
But any continuous value can be used...
Multi Dimensional Gene Set Analysis
==============================================================
Genomic dimensions
--------------------------------------------------------------
- Gene expression
- Transcript expression
- Splicing
- microRNA activity
- Promoters
- Methylation
- SNV (genetic variation)
- Copy Number
- Loss of heterozygosity (LOH)
--------------------------------------------------------------
- Study several diseases / datasets at the same time
- Gene conservation in evolution
- Text mining information: genes vs. treatments
- Some other physical characteristics: gene length
--------------------------------------------------------------
![](images/GO_xl.png)
Multidimensional Logistic Regression
--------------------------------------------------------------
\
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha\ x_1 + \beta\ x_2 $$
--------------------------------------------------------------
![](images/figure1.png)
Interaction
--------------------------------------------------------------
\
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha\ x_1 + \beta\ x_2 + \gamma\ x_1 x_2 $$
--------------------------------------------------------------
![](images/figure2a.png)
--------------------------------------------------------------
![](images/figure2b.png)
--------------------------------------------------------------
![](images/GO_bimodal2.png)
--------------------------------------------------------------
![](images/GO_interaccion.png)
Interpretation
--------------------------------------------------------------
\
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha\ x_1 + \beta\ x_2 + \gamma\ x_1 x_2 $$
--------------------------------------------------------------
![](images/interpret_1.png)
--------------------------------------------------------------
![](images/interpret_2.png)
--------------------------------------------------------------
![](images/interpret_3.png)
Extensions
==============================================================
microRNAs
--------------------------------------------------------------
![](images/mirna.png)
Target genes
--------------------------------------------------------------
![](images/mirna_targets.png)
Biological function
--------------------------------------------------------------
```
Function (GO)
/ \
+ gene expression -> + protein -> + biological activity
^
|
+ miRNA expression -> - protein -> - biological activity
\ /
Function (GO)
```
- miRNA protein inhibition through _target_ genes
- Functionality is _annotated_ to genes: GO db
- miRNAs _target_ certain genes: miRTarBase db
All information converges at a gene level
where the functional interpretation can be done.
microRNA to gene transference
--------------------------------------------------------------
$$ t_i = \sum_{j \in GT_i} r_{j} $$
- $r_{j}$ ranking statistic for miRNA $j$
- $GT_i$ set of miRNAs targeting gene $i$
- $t_{i}$ ranking statistic for gene $i$: __inhibition score__
Additive (or subtractive) score: simple but efficient model
Plug into the logistic model
--------------------------------------------------------------
\
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha\ x_1 + \beta\ x_2 + \gamma\ x_1 x_2 $$
\
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha\ t + \beta\ x_2 + \gamma\ t x_2 $$
Unidimensional case
--------------------------------------------------------------
\
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha \ r_i $$
\
$$ \log \frac{P(g_i \in GO)}{P (g_i \notin GO)} = \kappa + \alpha \ t_i $$
--------------------------------------------------------------
![](images/diagram.png)
--------------------------------------------------------------
![](images/interp_gen.png)
![](images/interp_go.png)
Some more Extensions
==============================================================
--------------------------------------------------------------
- Weighting methods: gene importance
- Gene to GeneSet empirical distance
- Gene set internal Correlation
- Meta analysis: combining results from several studies
- Supervised Classification using enriched GeneSets
- Unsupervised Classification
- Much more work in data preprocessing and validation
- Account for technological biases: gene length NGS
- Organization and presentation of results:
thousands of tables, images, plots
And so...
==============================================================
It was fun
--------------------------------------------------------------
... and useful
- New methodology and tools are available
- I did my PhD
- I supervised one more PhD and several MSc dissertations
- Teaching and communication
[www.dmontaner.com/#Short Bioinformatics Courses](http://www.dmontaner.com/#Short%20Bioinformatics%20Courses)
+50 Publications _(peer review)_
--------------------------------------------------------------
- _Integrated gene set analysis for microRNA studies._ Garcia-Garcia F, Panadero J, Dopazo J, **Montaner D**. [Bioinformatics (Oxford, England). 2016; ](http://www.ncbi.nlm.nih.gov//pubmed/27324197)
- _Multidimensional gene set analysis of genomic data._ **Montaner D**, Dopazo J. [PloS one. 2010; 5(4):e10348.](http://www.ncbi.nlm.nih.gov//pubmed/20436964)
- _Gene set internal coherence in the context of functional profiling._ **Montaner D**, Minguez P, Al-Shahrour F, Dopazo J. [BMC genomics. 2009; 10:197.](http://www.ncbi.nlm.nih.gov//pubmed/19397819)
--------------------------------------------------------------
- _Babelomics 5.0: functional interpretation for new generations of genomic data._ Alonso R, et al. [Nucleic acids research. 2015; 43(W1):W117-21.](http://www.ncbi.nlm.nih.gov//pubmed/25897133)
- _Prophet, a web-based tool for class prediction using microarray data._ Medina I, **Montaner D**, Tárraga J, Dopazo J. [Bioinformatics (Oxford, England). 2007; 23(3):390-1.](http://www.ncbi.nlm.nih.gov//pubmed/17138587)
- _Next station in microarray data analysis: GEPAS._ **Montaner D**, et al. [Nucleic acids research. 2006; 34(Web Server issue):W486-91.](http://www.ncbi.nlm.nih.gov//pubmed/16845056)
[www.dmontaner.com/#Papers](http://www.dmontaner.com/#Papers)
Software
--------------------------------------------------------------
- http://bioconductor.org/packages/mdgsa
- http://bioconductor.org/packages/SNPediaR
- https://github.com/dmontaner/mirbaseID
- https://github.com/genometra/agilent
- https://cran.r-project.org/web/packages/TiddlyWikiR
[wiki example](http://dmontaner-papers.github.io/gsa4mirna)
http://babelomics.org
Knowledge Transfer
--------------------------------------------------------------
![](images/logo_genometra.jpg)
[www.genometra.com](http://www.genometra.com)
First [CIPF](https://www.cipf.es/web/portada/spinoffs;jsessionid=C2DE665D4DFEE112095BB7B311E1A7A3) Spin-off Company
& ? 2 u
==============================================================
--------------------------------------------------------------
- Statistics & Machine Learning
- Theory + applied skills + statistical intuition
- Wide computational skills
- Problem solver: beyond data analysis
- Problem finder: new questions
- Team building
- Training and teaching (colleagues and myself)
- Reporting and communicating
- Collaborating (open source)
- Risk taking...
Risk is Risk...
--------------------------------------------------------------
... in Finances or Biology
Thank you!
--------------------------------------------------------------