-
Notifications
You must be signed in to change notification settings - Fork 0
/
how-to-talk-to-your-bioinformatician.qmd
1204 lines (814 loc) · 26.6 KB
/
how-to-talk-to-your-bioinformatician.qmd
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
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
---
title: "How to talk to your bioinformatician?"
author:
- name: "January Weiner"
orcid: 0000-0003-1438-7819
affiliations:
- ref: cubi
affiliations:
- id: cubi
name: Core Unit for Bioinformatics, BIH@Charité
address: Charitéplatz 1,
postal-code: 10117
city: Berlin
country: Germany
title-slide-attributes:
data-background-image: files/bih_bg_logo.png
format:
revealjs:
footer: "Core Unit for Bioinformatics, BIH@Charite"
theme: cubi.scss
logo: files/bih_logo_small.png
transition: fade
slide-number: "c/t"
smaller: true
navigation-mode: linear
self-contained: true
comments:
hypothesis: true
knitr:
opts_chunk:
dev: "svg"
---
## About this presentation
Chances are, you have opened a PDF version of this presentation. Better
go to the [HTML version at https://bihealth.github.io/howtotalk](https://bihealth.github.io/howtotalk) to see the
original layout and the newest version.
# Who am I to tell you things?
```{scss}
$callout-color-tip: #70ADC1;
```
```{css, echo=FALSE}
.large {
font-size: 1.5em !important;
background-color: "#481567" !important;
color: white !important;
}
.theme-light .mermaid {
--color-red: #ff7777;
--color-orange: #ffbb99;
--color-yellow: #ffeeaa;
--color-green: #bbffbb;
--color-cyan: #bbffee;
--color-blue: #aaccff;
--color-purple: #ddbbff;
--color-pink: #ffcccc;
}
.viri1 {
background-color: "#481567";
color: white;
}
.whosaid {
text-align: right !important;
}
.aside {
font-size: 0.65em !important;
}
```
```{r echo=FALSE,warning=FALSE,message=FALSE}
#| label: setup
library(tidyverse)
library(ggplot2)
theme_set(theme_minimal())
```
## Bioinformatics, statistics, computational biology
```{dot}
#| fig-width: 12
graph G {
layout = "neato"
fontname = "Helvetica,Arial,sans-serif"
scale = 1.2
node [
style = filled
color = "#000000"
fontname = "Helvetica,Arial,sans-serif"
fontsize = 10
shape = "box"
]
statistics [ fillcolor="#F8B229", style=filled, label="Statistics", fontsize=18 ]
bioinformatics [ fontcolor="white", fillcolor="#481567", style=filled, label="Bioinformatics", fontsize=18 ]
cbio [ fontcolor="white", fillcolor="#33638d", style=filled, label="Computational\nbiology", fontsize=18 ]
"Computers and science" [ fontcolor="white", fontsize=20, label = "Computers\nand\nScience" ]
"Computers and science" -- bioinformatics
"Computers and science" -- cbio
"Computers and science" -- statistics
subgraph Bioinformatics {
node [ style=filled, fillcolor="#481567cc", fontcolor="white", fontsize=14 ]
bioinformatics -- algorithms -- implementation
bioinformatics -- formalizations
bioinformatics -- standards
bioinformatics -- databases -- maintenance
}
subgraph cbio {
node [ fontcolor="white", fillcolor="#33638dcc", style=filled, fontsize=14 ]
cbio -- "data\nscience" -- "data\nstewardship"
cbio -- "functional\ngenomics"
"functional\ngenomics" -- "gene set\nenrichment"
cbio -- "using tools"
cbio -- "sequence\nanalysis" -- "phylogenies"
}
subgraph Statistics {
node [ fillcolor="#F8B22999cc", style=filled, fontsize=14 ]
statistics -- "hypothesis\ntesting"
"hypothesis\ntesting" -- "p-values"
"hypothesis\ntesting" -- "effect sizes"
statistics -- "power analysis"
statistics -- "linear\nmodeling" -- "hierarchical\nmodeling"
statistics -- "visualizations"
statistics -- "Bayesian\nstatistics"
}
}
```
::: {.notes}
* What you call "bioinformatics" are different areas of science
* We can't know all biology, so be patient with us
* Bioinformatics is not a service
* Bioinformatics is a scientific collaboration
Be patient with us. We are not stupid, but we don't know your context, your
field, your abbreviations.
:::
##
::: {.callout-tip}
## First bottom line
**Talk!**
Keep explaining your project – teach us!
(it goes both ways)
:::
## Things bioinformaticians care about {.incremental}
* The biological question
* Statistics
* Experimental design
* Quality control
* Reproducibility
* Consistency
::: {.notes}
Differences between "amateur" bioinformaticians like me a while ago and
"professionals" are mostly in the QC and reproducibility part
:::
# Statistics
## Statistics matters
![](images/statistics.png)
## What is a p-value?
$H_0$: The null hypothesis, no effect
$H_1$: The alternative hypothesis, there is an effect
We run a test, we get a p-value, say $0.03$. It is a probability.
Probability of *what*, exactly?
Raise your hands if you think that the p-value is the...
. . .
1. Probability that $H_0$ is true, given the data
. . .
1. Probability that $H_1$ is wrong, given the data
. . .
1. Probability that the data is random
. . .
1. Probability that the observations are due to random chance
. . .
1. Probability of getting the same data by random chance
. . .
* **Probability of observing an effect at least as extreme given that $H_0$ is true**
::: {.notes}
None of that is true.
P-values are not what we intiutively think they are.
They are in fact way less useful than we think they are.
The $H_0$ is true or false, there is no probability involved.
:::
## Our intuition is bayesian, not frequentist
| **Frequentist Statistics** | **Bayesian Statistics** |
|-------------------------------|-----------------------------|
| ![](images/classicalgiants.png)| ![](images/thomas-bayes.png) |
| 1. Probability is defined as the long-run frequency of events | 1. Probability represents a degree of belief or certainty about an event |
| 2. Parameters (like the "true value") are fixed but unknown quantities. | 2. Parameters are treated as random variables with their own probability distributions. |
| 3. Asking about the probability of a hypothesis does not make sense | 3. Asking about the probability of a hypothesis is the main goal |
::: {.aside}
Ask me later to show you the coin trick!
:::
## Why is that important?
P-values are the *language* of science, whether we like them (we don't) or
not.
* *Always* use effect sizes
* Never rely on p-values alone
* Know their limits!
::: {.callout-tip}
## P-values are a language
You have to understand p-values and their limits to talk to other
scientists!
:::
::: {.aside}
Wasserstein RL, Lazar NA. The ASA statement on p-values: context, process, and purpose. The American Statistician. 2016 Apr 2;70(2):129-33.
:::
## How Venn diagrams can fool scientists
COVID-19 study, both COVID-19 patients and non-COVID-19 patients are
compared in two groups of people, *G1* and *G2*.
We wanted to know whether the influence of COVID-19 is different in these
two groups.
![](images/fig1-1.jpeg)
::: {.aside}
Venn diagrams may indicate erroneous statistical reasoning in
transcriptomics. Weiner, Obermayer and Beule, 2022, Frontiers Genetics
:::
## What happens is, we are comparing significance with non-significance
> The Difference Between “Significant” and “Not Significant” is not
> Itself Statistically Significant
::: {.whosaid}
(*Andrew Gelman and Howard Stern*)
:::
If a gene is significant in one comparison, and not significant in another,
that does not mean that there is a difference between the two groups.
It simply means that we *failed* to detect the difference in one of the
comparisons, but that is actually quite likely to happen!
::: {.callout-tip}
## Therefore:
Don't say "there is no difference". Say "we did not detect a difference".
:::
::: {.aside}
Gelman A, Stern H. The difference between “significant” and “not significant” is not itself statistically significant. The American Statistician. 2006 Nov 1;60(4):328-31.
:::
## How Venn diagrams can fool scientists
Nieuwenhuis et al. found that half of the scientists who could have
commited this error, did in fact commit this error.
![](images/erroneous_analyses.webp)
::: {.aside}
Nieuwenhuis, Sander, Birte U. Forstmann, and Eric-Jan
Wagenmakers. "Erroneous analyses of interactions in neuroscience: a problem
of significance." Nature neuroscience 14.9 (2011): 1105-1107.
:::
## The results are artifacts!
Groups G1 and G2 were randomly drawn from the same population. They were
not different at all.
![](images/fig2-1.jpeg)
## Going beyond p-values
:::: {.columns}
::: {.column width="45%"}
* Estimation rather than testing (e.g. confidence intervals rather than
p-values)
* Considering effect sizes
* Power analysis – estimating type II error rates (false negatives)
* Sign / magnitude errors
* Bayesian statistics
* Correcting for multiple testing
:::
::: {.column width="10%"}
:::
::: {.column width="45%"}
![](images/pregnant.jpg)
:::
::::
# Experimental design
## How many samples are sufficient?
Say, we want to compare two groups with a standard $t-test$, nothing fancy.
Our ability to detect the differences (the statistical *power*) depends on
the sample size and the effect size[^cohen].
```{r echo=FALSE}
#| label: effect_sizes
#| fig-width: 9
#| fig-height: 4
# plot four effect sizes as boxplots
library(tidyverse)
library(ggplot2)
library(ggbeeswarm)
n_samp <- c(3, 4, 5, 7, 10, 15, 25)
d <- c(Large=.8, "Very large"=1.2, Huge=2.0, Ludicrous=4.0)
n_samples <- 100
df <- imap_dfr(d, ~ {
d_val <- .x
size <- .y
x <- rnorm(n_samples, mean = 0, sd = 1)
y <- rnorm(n_samples, mean = d_val, sd = 1)
data.frame(value = c(x, y), group = rep(c("A", "B"), each = n_samples), d = d_val, size=size)
}) %>%
mutate(size = paste0(size, " (d=", d, ")")) %>%
mutate(size = factor(size, levels=c("Large (d=0.8)", "Very large (d=1.2)", "Huge (d=2)", "Ludicrous (d=4)")))
ggplot(df, aes(x = group, y = value, fill = group)) +
geom_boxplot(outlier.shape = NA) +
geom_beeswarm(alpha=.25, corral.width=1.5) +
facet_wrap(~ size, ncol=4) +
labs(x = "Group", y = "Value", fill = "Group") +
theme(legend.position = "none")
```
[^cohen]: Here we use Cohen's $d$
## How many samples are sufficient?
The $y$ axis on this plot shows how the power of the test – meaning how
often, assuming that the groups really differ by $d$ on average, you will
be able to detect the difference using a t-test.
```{r echo=FALSE,cache=TRUE}
#| label: power
# we could use pwr package here, but simulation is more explicit
# parameters
n <- 10000
# simulation
power <- map_dfr(n_samp, ~ {
n_samples <- .x
map_dfr(d, ~ {
d_val <- .x
pwr <- map_dbl(1:n, ~ {
x <- rnorm(n_samples, mean = 0, sd = 1)
y <- rnorm(n_samples, mean = d_val, sd = 1)
t.test(x, y)$p.value
}) < 0.05
data.frame(n_samples = n_samples, d = d_val, power = mean(pwr))
})
}) %>%
mutate(study = "Simple t-test")
```
```{r}
#| label: power_plot_simple
#| fig-width: 9
#| fig-height: 4
power$n_samples_tot <- 2 * power$n_samples
ggplot(power, aes(x = n_samples, y = power, color = factor(d))) +
geom_point() +
geom_line() +
scale_color_brewer(palette = "Dark2", name="Effect size",
guide=guide_legend(reverse=TRUE)) +
scale_x_log10() +
labs(x = "Number of samples (per group)", y = "Power")
```
**Power**: power of 50% means that on average, you will be able to see
statistical significance in 50% of the experiments *assuming there is a
difference!*
## How many samples are sufficient?
What about the following setup:
* We have 2 strains (WT and KO)
* We have treatment + control
* We want to know whether the treatment has a different effect on the KO
strain than on the WT strain
This is a 2x2 design, and we need to consider the interaction term.
## How many samples are sufficient?
```{r echo=FALSE,cache=TRUE}
#| label: power_int
# we could use pwr package here, but simulation is more explicit
library(broom)
# parameters
n <- 1000
# simulation
power_int <- map_dfr(n_samp, ~ {
n_samples <- .x
map_dfr(d, ~ {
d_val <- .x
pwr <- map_dbl(1:n, ~ {
x <- rnorm(n_samples, mean = 0, sd = 1)
y <- rnorm(n_samples, mean = 0, sd = 1)
z <- rnorm(n_samples, mean = 0, sd = 1)
w <- rnorm(n_samples, mean = 1 * d_val, sd = 1)
df <- data.frame(x=c(x, y, z, w),
strain=rep(c("WT", "KO"), each=2 * n_samples),
group=rep(c("Control", "Treatment"), each=n_samples))
mod <- lm(x ~ strain * group, data = df)
anova(mod)$`Pr(>F)`[3]
}) < 0.05
data.frame(n_samples = n_samples, d = d_val, power = mean(pwr))
})
}) %>%
mutate(study = "Interaction")
```
```{r echo=FALSE}
#| label: power_plot
power_int$n_samples_tot <- power_int$n_samples * 4
power_tot <- rbind(power, power_int) %>%
mutate(study = factor(study, levels=c("Simple t-test", "Interaction")))
power_tot$dfac <- factor(power_tot$d)
power_tot$dfac <- factor(power_tot$d, levels=rev(levels(power_tot$dfac)))
ggplot(power_tot, aes(x = n_samples, y = power, color = factor(d))) +
geom_point() +
geom_line() +
scale_color_brewer(palette = "Dark2", name="Effect size",
guide=guide_legend(reverse=TRUE)) +
labs(x = "Number of samples (per group)", y = "Power") +
facet_wrap(~ study, scales = "free_x")
```
## How many samples are sufficient?
That is not even the worse thing.
Simple calculations show that assuming
* your power is 80% (really great!)
* $p-value$ cutoff is $0.05$
* 90% of the $H_0$ are true (i.e., 10% of the time the differences are
real)
then 36% of your "significant" results **are false positives**[^more]!
(Plus, you failed to detect 20% of the real differences)
[^more]: I can walk you through a very simple demonstration later if you
care.
::: {.aside}
Colquhoun D. An investigation of the false discovery rate and the misinterpretation of p-values. Royal Society open science. 2014 Nov 19;1(3):140216.
:::
##
::: {.callout-tip}
## Bottom line
Talk to your statistician early!
Keep your study design simple!
:::
# Is high throughput data worth it?
## {.inverse background-color="#000000"}
:::: {.columns}
::: {.column width="45%"}
![](images/arnolfini.jpg)
:::
::: {.column width="10%"}
:::
::: {.column width="45%"}
::: {.r-stack}
![](images/arnolfini_2.jpg){.fragment}
![](images/arnolfini_3a.png){.fragment}
![](images/arnolfini_3.png){.fragment}
![](images/arnolfini_4.jpg){.fragment}
![](images/arnolfini_5.png){.fragment}
:::
:::
::::
::: {.notes}
:::
## {.inverse background-color="#000000"}
::: {.r-stack}
![](images/monet_4.jpg){.fragment width="800px"}
![](images/monet_3.jpg){.fragment width="800px"}
![](images/monet_2.jpg){.fragment width="900px"}
![](images/monet.jpg){.fragment width="1000px"}
:::
## {.inverse background-color="#000000"}
:::: {.columns}
::: {.column width="45%"}
![](images/monet.jpg)
:::
::: {.column width="10%"}
:::
::: {.column width="45%"}
![](images/sangiorgio.jpg)
:::
::::
## {.inverse background-color="#000000"}
![](images/monet_all.svg)
::: {.notes}
Claude Monet, Saint-Georges-Majeur, 1908, Venice – 37 paintings between 1.10 and 7.12
Almost as many as van Eyck painted during his whole life
:::
## Explorative vs hypothesis testing
:::: {.columns}
::: {.column width="45%"}
#### Explorative analysis
**Pro:**
* No need to define a-priori hypotheses
* Something unexpected and new can be found
* Can be used to generate hypotheses
**Con:**
* Requires multiple testing correction
* Requires proper validation
* Can't do it as the last step
:::
::: {.column width="10%"}
:::
::: {.column width="45%"}
#### Hypothesis-driven analysis
**Pro:**
* Clear questions
* Clear answers
* More statistical power
* Better story, better paper
**Con:**
* Requires more planning (and thinking!)
* Can make you miss something unexpected
* If you reject the hypothesis, tough luck
:::
::::
## The bottom line
:::: {.columns}
::: {.column width="45%"}
::: {.callout-tip}
## Do
* Formulate clear questions
* Manage your expectations
* Evaluate existing data - is the approach able to answer your questions?
* Read papers – which ones are similar to your study?
* Validate your results
:::
:::
::: {.column width="10%"}
:::
::: {.column width="45%"}
::: {.callout-warning .fragment}
## Don't
* expect miracles
* "let's just see what we can find"
* try to save money
* make too complex designs
:::
:::
::::
# Reproducibility
## Tale of two papers
![](images/pnas_1.png)
## Tale of two papers
![](images/pnas_res_1.png)
## Tale of two papers
![](images/pnas_1.png)
. . .
![](images/pnas_2.png)
## Tale of two papers
![](images/pnas_1.png)
![](images/pnas_2_red.png)
## Tale of two papers
![](images/pnas_res_2.png)
::: {.notes}
Key difference - second paper specifically focused on genes that were
regulated in either mouse model or human data, and asked whether they are
similar or not
:::
## Lessons learned
* *A lot* depends on how you analyze your data
. . .
* This in turn depends on the questions you ask
. . .
* The average "Methods" section is not sufficient for reproducible
science!
> " Second, **none of the 193 experiments** were described in sufficient detail
> in the original paper to enable us to design protocols to repeat the
> experiments, so we had to seek clarifications from the original authors."
> (Errington et al., 2021)
::: {.aside}
Errington TM, Mathur M, Soderberg CK, Denis A, Perfito N, Iorns E, Nosek BA. Investigating the replicability of preclinical cancer biology. Elife. 2021 Dec 10;10:e71601.
Errington TM, Denis A, Perfito N, Iorns E, Nosek BA. Challenges for assessing replicability in preclinical cancer biology. Elife. 2021 Dec 7;10:e67995.
:::
# File formats and data management
## How we work
```{mermaid}
flowchart LR
A(Excel) --> B(Data import)
AA(CSV, TSV) --> B(Data import)
AAA(fastq, ...) --> B(Data import)
AAAA("Biological\nquestion") --> B(Data import)
B --> C[Data\ncleanup]
C --> CC[QC]
CC --> E[Analysis]
E --> D
E --> F(Figures)
E --> G(Manuscript\nfragments)
E --> H(Tables\nExcel files)
F --> I[You]
G --> I
H --> I
I --> J[Fine tuning]
J --> E
C --> D[(Long term storage)]
```
## How we work
```{mermaid}
flowchart LR
A(Excel) --> B(Data import)
AA(CSV, TSV) --> B(Data import)
AAA(fastq, ...) --> B(Data import)
AAAA("Biological\nquestion") --> B(Data import)
B --> C[Data\ncleanup]
C --> CC[QC]
CC --> E[Analysis]
E --> D
E --> F(Figures)
E --> G(Manuscript\nfragments)
E --> H(Tables\nExcel files)
F --> I[You]
G --> I
H --> I
I --> J[Fine tuning]
J --> E
C --> D[(Long term storage)]
style C color:#000,fill:#f9f,stroke:#333,stroke-width:4px
style J color:#000,fill:#f9f,stroke:#333,stroke-width:4px
```
In the diagram above, two things take usually a lot of hands-on time:
* Understanding and cleaning the data
* Fine-tuning the analysis results
## Excel and gene names
:::: {.columns}
::: {.column width="60%"}
* Excel converts some words to dates automatically
* Gene names like `MARCH1` or `SEPT9` are converted to dates
* In most cases[^cases], you can't switch off this behavior
[^cases]: You can change the data type for a column to "text" before
pasting data in, but this is just a workaround. In Office 365 it is
possible to switch off this behavior.
:::
::: {.column width="5%"}
:::
::: {.column width="35%"}
:::
::::
## Excel and gene names
::: {.r-stack}
![](images/excel_paper_1.png){.fragment .absolute top=90 left=10 width=80%}
![](images/excel_hgnc.png){.fragment .absolute top=120 left=50 width=80%}
![](images/excel_paper_2.png){.fragment .absolute top=220 left=90 width=80%}
:::
::: {.aside style="font-size: 0.8em; color:red;"}
::: {style="font-size: 0.5em;"}
Ziemann, Mark, Yotam Eren, and Assam El-Osta. "Gene name errors are widespread in the scientific literature." Genome biology 17 (2016): 1-3;
Abeysooriya M, Soria M, Kasu MS, Ziemann M. Gene name errors: Lessons not learned. PLoS Computational Biology. 2021 Jul 30;17(7):e1008984.
:::
:::
::: {.notes}
2016 paper: 20% of papers have Excel gene name errors
2021 paper: it is 30% now. Excel now learned mulitple languages
:::
## Is Excel suitable for science?
* How do you record changes?
* How do you prevent automatic changes?
* In short – how do you ensure reproducibility?
## How to give us (meta-)data
Part of the communication is passing on the data.
1. Make sure the data is **complete** (batches? replicates?)
1. Identifiers should be unique and non-numeric (`ID1` rather than `1`)
2. Use a separate sheet to describe the meaning of columns
3. Explain abbreviations
4. Make the data machine-friendly
5. Disclose precisely all methods (like models, kit labels etc, request them from service providers!)
## What you should demand from your bioinformaticians
* Methods used
* Scripts / pipelines used
* Full processed data
* All results as tables (Excel, CSV etc)
Even if you don't know what to do with all that *now*, you might be needing
it in the future!
## How (not to) work with Excel
(for your reference)
* Avoid manually changing Excel files
* Never use formatting for data
* Don't combine values and comments
* Don't put meta-information into column names
* One sheet = one table
* Header = one line
* Do not use merged cells
* Use consistent file names
* Avoid spaces in file and column names (use underscores)
# Some more tips and summaries
##
:::: {.columns}
::: {.column width="45%"}
### Things we don't like
* Cleaning up data
* Data dredging
* P-hacking
* Post-hoc hypotheses
* Excel
* Manual changes like changing fonts in figures
* Non-reproducible science
:::
::: {.column width="10%"}
:::
::: {.column width="45%"}
### Things we love
* Clear questions
* A priori hypotheses
* Challenging statistics
* Creating new tools
* R and Rmarkdown, or
* Python and Jupyter
* Reproducible workflows
* Well organized data
:::
::::
## Things that you might want to learn
* Statistics
* Coding (likely R or Python)
* Reproducible workflows with Quarto/Rmarkdown or Jupyter
Even if you are not going to use these tools yourself, gaining an insight
into how they work will help you to communicate with your bioinformatician.
## Thank you {.inverse background-color="#000000"}
:::: {.columns}
::: {.column width="40%"}
You can find this presentation along its source code
at [https://github.com/bihealth/howtotalk](https://github.com/bihealth/howtotalk)
A 5 day R crash course book is available at
[https://bihealth.github.io/RCrashcourse-book/](https://bihealth.github.io/RCrashcourse-book/)
Course materials & videos:
[https://bihealth.github.io/RCrashcourse2023/](https://bihealth.github.io/RCrashcourse-book/)
:::
::: {.column width="10%"}
:::
::: {.column width="50%"}
```{r}
#| label: qr
#| fig-width: 4.5