-
Notifications
You must be signed in to change notification settings - Fork 0
/
random_QC_samples.Rmd
1128 lines (979 loc) · 44.7 KB
/
random_QC_samples.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
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: "Evaluation data quality of QC samples from 2020-09"
author: "Johannes Rainer, Vinicius Veri"
output:
BiocStyle::html_document:
toc: true
number_sections: false
toc_float: true
bibliography: references.bib
csl: biomed-central.csl
---
```{r biocstyle, echo = FALSE, results = "asis" }
library(BiocStyle)
BiocStyle::markdown()
knitr::opts_chunk$set(message = FALSE, warning = FALSE)
```
**Modified**: `r file.info("random_QC_samples.Rmd")$mtime`<br />
**Compiled**: `r date()`
```{r settings, echo = TRUE, results = "hide", message = FALSE}
## Set general options
options(useFancyQuotes = FALSE)
set.seed(123)
## Define paths:
filename <- "random_QC_samples"
## Path to save the images; remove all old images.
IMAGE_PATH <- paste0("images/", filename, "/")
dir.create(IMAGE_PATH, recursive = TRUE, showWarnings = FALSE)
## Path to store RData files
RDATA_PATH <- paste0("data/RData/", filename, "/")
dir.create(RDATA_PATH, recursive = TRUE, showWarnings = FALSE)
## Get the number of cpus allocated or fall back to 6
ncores <- as.integer(Sys.getenv("SLURM_JOB_CPUS_PER_NODE", 6))
rt_cut <- 340
MZML_PATH <- "/data/massspec/mzML/"
if (!dir.exists(MZML_PATH))
stop("Can not find the directory with the mzML files: ", MZML_PATH)
```
# Introduction
After the Covid-19 caused lockdown the LC-MS system seemed to no longer generate
reproducible data. In this document we evaluate the signal measured for QC
samples in September 2020 (after thorough revision and cleaning but prior
maintenance of the LC-MS system) and compare that signal to the data of random
QC samples measured in the year 2017. In particular, we aim at comparing the TIC
and BPC of all samples and evaluate the retention time and integrated peak
signal for selected ions.
```{r libraries, message = FALSE}
library(readxl)
library(RColorBrewer)
library(xcms)
library(CompMetaboTools)
library(Spectra)
injections <- as.data.frame(read_xlsx("data/random-QC-2017-2020_09.xlsx"))
injections$id <- gsub(".mzML", "", injections$mzML_file)
injections <- injections[order(injections$timestamp), ]
injections$batch <- injections$run_id
injections <- injections[!injections$type %in% c("ACN_IS", "EQ_MIX"), ]
## split by polarity
pd_pos <- injections[injections$polarity == "POS", ]
```
Below we load the positive polarity data, subset to the defined retention time
range and remove unnecessary spectrum metadata.
```{r pos-load, eval = !file.exists(paste0(RDATA_PATH, "data_pos.RData"))}
data_pos <- readMSData(files = paste0(MZML_PATH, pd_pos$mzML_file),
pdata = as(AnnotatedDataFrame(pd_pos),
"NAnnotatedDataFrame"),
mode = "onDisk")
data_pos <- filterRt(data_pos, c(0, rt_cut))
data_pos <- selectFeatureData(
data_pos, fcol = c("fileIdx", "spIdx", "acquisitionNum", "retentionTime",
"msLevel", "precursorScanNum", "polarity", "centroided",
"seqNum")
)
```
# Peak detection and refinement
We perform a centWave-based peak detection followed by a *peak refinement* step
that aims at reducing the peak detection artifacts, such as split peaks or
overlapping peaks. Below we define the parameters for the peak detection and
subsequently perform the analysis.
```{r pos-peak-detection, warning = FALSE, message = FALSE, eval = !file.exists(paste0(RDATA_PATH, "data_pos.RData"))}
cwp <- CentWaveParam(
peakwidth = c(2, 20),
ppm = 50,
snthresh = 5,
mzdiff = 0.001,
prefilter = c(3, 500),
noise = 100,
integrate = 2)
register(MulticoreParam(ncores))
tm <- system.time(
data_pos <- findChromPeaks(data_pos, param = cwp)
)
save(data_pos, file = paste0(RDATA_PATH, "data_pos.RData"))
```
```{r pos-peak-detection-load, echo = FALSE, warning = FALSE, message = FALSE}
load(paste0(RDATA_PATH, "data_pos.RData"))
if (!any(ls() == "tm"))
tm <- c(elapsed = 1)
```
Next we refine the identified chromatographic peaks. Processing took
`r tm["elapsed"] / 60` minutes.
```{r pos-peak-refinement, warning = FALSE, message = FALSE, eval = !file.exists(paste0(RDATA_PATH, "data_pos_ref.RData"))}
mnp <- MergeNeighboringPeaksParam(
expandRt = 2,
expandMz = 0.001,
ppm = 10,
minProp = 0.66)
## register(MulticoreParam(workers = ceiling(ncores * 2/3),
## timeout = 86400L))
tm <- system.time(
data_pos <- refineChromPeaks(data_pos, param = mnp, BPPARAM = bpparam())
)
save(data_pos, file = paste0(RDATA_PATH, "data_pos_ref.RData"))
```
```{r pos-peak-refinement-load, echo = FALSE, warning = FALSE, message = FALSE}
load(paste0(RDATA_PATH, "data_pos_ref.RData"))
if (!any(ls() == "tm"))
tm <- c(elapsed = 1)
```
```{r echo = FALSE, eval = FALSE}
## Just testing the joyPlot...
plotChromPeaks(data_pos, file = 12)
abline(v = c(150, 174))
a <- filterFile(data_pos, file = 12)
b <- filterFile(data_pos, file = 2)
plot(chromatogram(a, include = "none"))
abline(v = c(150, 174))
par(mfrow = c(1, 2))
## Focusing on large-mass peaks
joyPlot(b, rt = c(150, 174), mz = c(400, 1000), yoffset = 0.7)
joyPlot(a, rt = c(150, 174), mz = c(400, 1000), yoffset = 0.7)
plotChromPeaks(data_pos, file = 12)
abline(v = c(40, 60))
joyPlot(a, rt = c(40, 60), yoffset = 0.7, mz = c(200, 1000), spacing = "equal", legend = TRUE)
grid()
## Why the hell is this signal shifted in RT?
## "Cluster" compound already in solution -> different rt.
## Phosoholipid peak
plot(chromatogram(a, include = "none"))
abline(v = c(125, 143))
par(mfrow = c(1, 2))
joyPlot(b, rt = c(125, 143), yoffset = 0.7)
joyPlot(a, rt = c(125, 143), yoffset = 0.7)
## The *real* version...
par(bg = "#000000")
joyPlot(a, col = "#ffffff", peakBg = "#000000aa", mz = c(400, 1000),
yoffset = 0.8, rt = c(150, 174))
```
Processing took `r tm["elapsed"] / 60` minutes.
# Signal evaluation
## Base peak and total ion chromatograms
Next we extract the base peak and total ion chromatogram.
```{r chrs}
data_pos$year_month <- paste0(data_pos$year, "-", data_pos$month)
bpchr <- chromatogram(as(data_pos, "OnDiskMSnExp"), aggregationFun = "max")
tichr <- chromatogram(as(data_pos, "OnDiskMSnExp"), aggregationFun = "sum")
```
Below we plot the BPC and the TIC for the data set.
```{r chr-plots, fig.path = IMAGE_PATH, caption = "BPC and TIC of the data set", fig.width = 12, fig.height = 10, echo = FALSE}
col_year_month <- brewer.pal(8, "Set1")[-6]
names(col_year_month) <- unique(data_pos$year_month)
par(mfrow = c(2, 1), mar = c(2, 4.5, 2, 1))
plot(bpchr, col = paste0(col_year_month[bpchr$year_month], 80), main = "BPC",
xlim = c(0, 220))
grid()
legend("topright", legend = names(col_year_month), lty = 1, col = col_year_month)
plot(tichr, col = paste0(col_year_month[tichr$year_month], 80), main = "TIC",
xlim = c(0, 220))
grid()
```
The signal in September 2020 seems to start and stop earlier than in previous
years or also in February 2020. The bimodal *phospholipid* peak present in most
samples seems to be replaced by a single peak in some of the September samples,
with the exception of the last 5 samples which again show a large phospholipid
peak and a less noisy baseline.
Next we plot the distribution of the base peak and total ion signal per sample
to evaluate potential differences in average intensities.
```{r chr-boxplot, fig.path = IMAGE_PATH, caption = "Distribution of base peak and total ion chromatograms.", fig.width = 8, fig.height = 8, echo = FALSE}
par(mfrow = c(2, 1), mar = c(1, 4.5, 1, 1))
boxplot(lapply(bpchr, function(z) log2(intensity(z))), outline = FALSE,
ylab = expression(log[2]~base~peak~intensity), xaxt = "n",
main = "QC samples, positive polarity",
col = paste0(col_year_month[bpchr$year_month], 40))
grid(nx = NA, ny = NULL)
boxplot(lapply(tichr, function(z) log2(intensity(z))), outline = FALSE,
ylab = expression(log[2]~total~ion~intensity), xaxt = "n",
main = "QC samples, positive polarity",
col = paste0(col_year_month[tichr$year_month], 40))
grid(nx = NA, ny = NULL)
```
The distribution of base peak and total ion signal clearly shows batch
differences over the years, with the QC samples from the same batch yielding,
with the exception of the first 2 from the September 2020 (pink samples), about
the same signal. Also, the signal for the following 4 samples seems to
continuously decrease. This change is however comparably small and even in
previous batches such a difference was observed. The last 5 samples show almost
no such drift.
## Peak detection results
Next we evaluate the number of identified peaks per sample.
```{r peak-count-table, results = "asis", echo = FALSE}
library(pander)
T <- matrix(as.integer(table(chromPeaks(data_pos)[, "sample"])), ncol = 1)
colnames(T) <- "peak count"
rownames(T) <- data_pos$name
T <- data.frame(T, year = data_pos$year, month = data_pos$month,
day = data_pos$day, check.names = FALSE)
pandoc.table(T, style = "rmarkdown",
caption = "Number of peaks identified in each sample")
```
There are considerable differences in the number of detected peaks across the
batches.
At last we evaluate the signal for selected compounds (from the set of
standards) and compare their retention time across the batches as well as their
quantified signal. To this end we load first the data with the expected
retention time and ion for the set of standards.
```{r standards}
library(CompoundDb)
library(Rdisop)
std_info <- read.table(
"https://raw.githubusercontent.com/EuracBiomedicalResearch/lcms-standards/master/data/standards_dilution.txt",
sep = "\t", header = TRUE, as.is = TRUE)
std_info <- std_info[!is.na(std_info[, "POS"]), ]
rownames(std_info) <- 1:nrow(std_info)
std_info$mzneut = NA
std_info$mz_ion = NA
for (i in seq(nrow(std_info))) {
if (grepl("C", std_info$formula[i])){
std_info$mzneut[i] <- getMolecule(
as.character(std_info$formula[i]))$exactmass
} else {
std_info$mzneut[i] = as.numeric(std_info$formula[i])
}
## Calculate also the m/z
std_info$mz_ion[i] <- unlist(
mass2mz(std_info$mzneut[i],
adduct = as.character(
std_info[i, "POS"])))
}
```
We then extract the EIC for each standard, visually inspect all of the plots and
determine for which standards we have an unambiguous mapping to a single peak.
```{r pos-eic-plots, echo = TRUE}
## Two-step approach: define first the rt/mz region for each standard and
## Extract the chromatogram afterwards.
std_info_mz_rt <- matrix(
ncol = 4, nrow = nrow(std_info),
dimnames = list(std_info$name, c("mzmin", "mzmax", "rtmin", "rtmax")))
for (i in 1:nrow(std_info)) {
pks <- chromPeaks(data_pos, mz = std_info$mz_ion[i], ppm = 50)
if (nrow(pks)) {
rtdiff <- abs(pks[, "rt"] - std_info$RT[i])
keep <- which(rtdiff < 40)
if (length(keep)) {
pks <- pks[keep, , drop = FALSE]
mzr <- range(pks[, c("mzmin", "mzmax")])
rtr <- range(pks[, c("rtmin", "rtmax")], std_info$RT[i])
std_info_mz_rt[i, ] <- c(mzr, rtr)
}
}
}
std_info <- std_info[!is.na(std_info_mz_rt[, "mzmin"]), ]
std_info_mz_rt <- std_info_mz_rt[!is.na(std_info_mz_rt[, "mzmin"]), ]
## Extract the ion chromatogram for each standard
rtr <- std_info_mz_rt[, c("rtmin", "rtmax")]
rtr[, 1] <- rtr[, 1] - 10
rtr[, 2] <- rtr[, 2] + 10
std_info_chr <- chromatogram(
data_pos, mz = std_info_mz_rt[, c("mzmin", "mzmax")],
rt = rtr, aggregationFun = "max")
## Plot the extracted ion chromatograms
dr <- paste0(IMAGE_PATH, "peakdetection/")
dir.create(dr, recursive = TRUE, showWarnings = FALSE)
sample_colors <- col_year_month[as.character(data_pos$year_month)]
for (i in 1:nrow(std_info)) {
cmpname <- std_info$name[i]
filename <- paste0(dr, gsub("%", "p", cmpname, fixed = TRUE),
"_POS", ".png")
png(file = filename, width = 16, height = 8, units = "cm",
res = 300, pointsize = 4)
chr <- std_info_chr[i, ]
pks <- chromPeaks(chr)
plot(chr, col = "#00000040",
main = paste0(std_info$name[i], ": ",
format(mz(chr)[1], digits = 6), "-",
format(mz(chr)[2], digits = 6)),
peakCol = paste0(sample_colors[pks[, "column"]], 50),
peakBg = paste0(sample_colors[pks[, "column"]], 10))
## This is just to indicate identified peaks with a rectangle.
xcms:::.add_chromatogram_peaks(
chr, pks,
col = paste0(sample_colors[pks[, "column"]], 40),
bg = NA,
type = "rectangle")
abline(v = std_info$RT[i])
dev.off()
}
```
```{r standards-manual-selection, echo = FALSE}
std_info$rtmin <- NA_real_
std_info$rtmax <- NA_real_
## Add the min and max retention time for compounds we select for further
## analysis
std_info$rtmin[std_info$name == "3-Nitrotyrosine"] <- 143
std_info$rtmax[std_info$name == "3-Nitrotyrosine"] <- 170
std_info$rtmin[std_info$name == "ADMA"] <- 163
std_info$rtmax[std_info$name == "ADMA"] <- 190
std_info$rtmin[std_info$name == "Alanine"] <- 165
std_info$rtmax[std_info$name == "Alanine"] <- 180
std_info$rtmin[std_info$name == "Arginine"] <- 170
std_info$rtmax[std_info$name == "Arginine"] <- 208
std_info$rtmin[std_info$name == "Asparagine"] <- 177
std_info$rtmax[std_info$name == "Asparagine"] <- 200
std_info$rtmin[std_info$name == "Caffeine"] <- 32
std_info$rtmax[std_info$name == "Caffeine"] <- 40
std_info$rtmin[std_info$name == "Citrulline"] <- 170
std_info$rtmax[std_info$name == "Citrulline"] <- 200
std_info$rtmin[std_info$name == "Creatine"] <- 160
std_info$rtmax[std_info$name == "Creatine"] <- 190
std_info$rtmin[std_info$name == "Cystine"] <- 205
std_info$rtmax[std_info$name == "Cystine"] <- 225
std_info$rtmin[std_info$name == "Diethyl malonate"] <- 25
std_info$rtmax[std_info$name == "Diethyl malonate"] <- 35
std_info$rtmin[std_info$name == "Glutamine"] <- 170
std_info$rtmax[std_info$name == "Glutamine"] <- 200
std_info$rtmin[std_info$name == "Glyceraldehyde 2-phosphate"] <- 180
std_info$rtmax[std_info$name == "Glyceraldehyde 2-phosphate"] <- 225
std_info$rtmin[std_info$name == "Glycero-phosphocholine"] <- 185
std_info$rtmax[std_info$name == "Glycero-phosphocholine"] <- 200
std_info$rtmin[std_info$name == "Glycine"] <- 165
std_info$rtmax[std_info$name == "Glycine"] <- 180
std_info$rtmin[std_info$name == "Histidine"] <- 170
std_info$rtmax[std_info$name == "Histidine"] <- 205
std_info$rtmin[std_info$name == "Homocysteine"] <- 165
std_info$rtmax[std_info$name == "Homocysteine"] <- 180
std_info$rtmin[std_info$name == "Indoleacetic acid"] <- 32
std_info$rtmax[std_info$name == "Indoleacetic acid"] <- 40
std_info$rtmin[std_info$name == "Inosine"] <- 135
std_info$rtmax[std_info$name == "Inosine"] <- 155
std_info$rtmin[std_info$name == "L-Aspartic Acid"] <- 175
std_info$rtmax[std_info$name == "L-Aspartic Acid"] <- 195
std_info$rtmin[std_info$name == "L-Glutamic Acid"] <- 165
std_info$rtmax[std_info$name == "L-Glutamic Acid"] <- 190
std_info$rtmin[std_info$name == "Lysine"] <- 170
std_info$rtmax[std_info$name == "Lysine"] <- 205
std_info$rtmin[std_info$name == "Methioninesulfoxide"] <- 178
std_info$rtmax[std_info$name == "Methioninesulfoxide"] <- 193
std_info$rtmin[std_info$name == "N-Acetylornithine"] <- 170
std_info$rtmax[std_info$name == "N-Acetylornithine"] <- 210
std_info$rtmin[std_info$name == "Ornithine"] <- 175
std_info$rtmax[std_info$name == "Ornithine"] <- 205
std_info$rtmin[std_info$name == "Putrescine"] <- 155
std_info$rtmax[std_info$name == "Putrescine"] <- 170
std_info$rtmin[std_info$name == "Serine"] <- 175
std_info$rtmax[std_info$name == "Serine"] <- 190
std_info$rtmin[std_info$name == "Sphingosine"] <- 25
std_info$rtmax[std_info$name == "Sphingosine"] <- 40
std_info$rtmin[std_info$name == "Suberic Acid"] <- 30
std_info$rtmax[std_info$name == "Suberic Acid"] <- 40
std_info$rtmin[std_info$name == "Taurine"] <- 165
std_info$rtmax[std_info$name == "Taurine"] <- 180
std_info$rtmin[std_info$name == "Threonine"] <- 165
std_info$rtmax[std_info$name == "Threonine"] <- 190
std_info$rtmin[std_info$name == "Tyrosine"] <- 160
std_info$rtmax[std_info$name == "Tyrosine"] <- 180
std_info <- std_info[!is.na(std_info$rtmin), ]
## Re-create the m/z and rt range matrix
std_info_mz_rt <- matrix(
ncol = 4, nrow = nrow(std_info),
dimnames = list(std_info$name, c("mzmin", "mzmax", "rtmin", "rtmax")))
for (i in 1:nrow(std_info)) {
pks <- chromPeaks(data_pos, mz = std_info$mz_ion[i], ppm = 50,
rt = c(std_info$rtmin[i], std_info$rtmax[i]))
if (nrow(pks)) {
mzr <- range(pks[, c("mzmin", "mzmax")])
rtr <- range(pks[, c("rtmin", "rtmax")], std_info$RT[i])
rtr[1] <- rtr[1] - 1
rtr[2] <- rtr[2] + 1
std_info_mz_rt[i, ] <- c(mzr, rtr)
}
}
std_chr <- chromatogram(
data_pos, mz = std_info_mz_rt[, c("mzmin", "mzmax")],
rt = std_info_mz_rt[, c("rtmin", "rtmax")],
aggregationFun = "max")
fData(std_chr) <- cbind(fData(std_chr),
std_info[, c("name", "HMDB.code", "formula",
"POS", "RT")])
```
Next we select for each of the selected standards the largest peak signal in
each sample (and also the retention time for that peak).
```{r select-peak-signal}
std_rt <- matrix(ncol = ncol(std_chr), nrow = nrow(std_chr),
dimnames = dimnames(std_chr))
rownames(std_info) <- std_info$name
rownames(std_rt) <- rownames(std_info)
std_int <- std_rt
std_max <- std_int
for (i in 1:nrow(std_rt)) {
tmp <- std_chr[i, ]
tmp <- split.data.frame(chromPeaks(tmp), chromPeaks(tmp)[, "column"])
tmp <- lapply(tmp, function(z) z[which.max(z[, "maxo"]), ])
rts <- vapply(tmp, function(z) z["rt"], numeric(1))
ints <- vapply(tmp, function(z) z["into"], numeric(1))
maxs <- vapply(tmp, function(z) z["maxo"], numeric(1))
std_rt[i, names(rts)] <- rts
std_int[i, names(ints)] <- ints
std_max[i, names(maxs)] <- maxs
}
idx <- order(std_info$RT)
std_info <- std_info[idx, ]
std_chr <- std_chr[idx, ]
std_rt <- std_rt[idx, ]
std_int <- std_int[idx, ]
std_max <- std_max[idx, ]
```
The table below lists the manually selected standards which will be used for the
subsequent analysis.
```{r select-standards-table, echo = FALSE, results = "asis"}
T <- data.frame(std_info[, c("name", "mz_ion")],
rt_mean = apply(std_rt, 1, mean, na.rm = TRUE),
rt_sd = apply(std_rt, 1, sd, na.rm = TRUE),
into_mean = apply(std_int, 1, mean, na.rm = TRUE),
into_cv = rowRsd(std_int, na.rm = TRUE))
pandoc.table(T, style = "rmarkdown",
caption = "Selected standards.")
```
Retention time shifts for about half of the standards is relatively large (over
6 seconds). Coefficient of variation are large for all standards.
```{r export-eic-selected-standards, echo = FALSE, message = FALSE}
dr <- paste0(IMAGE_PATH, "selected-standards/")
dir.create(dr, recursive = TRUE, showWarnings = FALSE)
sample_colors <- col_year_month[as.character(data_pos$year_month)]
for (i in seq_len(nrow(std_chr))) {
fn <- paste0(dr, std_info$name[i], ".png")
png(file = fn, width = 16, height = 10, units = "cm",
res = 300, pointsize = 4)
par(mfrow = c(1, 2), mar = c(5, 4.2, 2, 0.5))
chr <- std_chr[i, ]
pks <- chromPeaks(chr)
plot(chr, col = "#00000040",
main = paste0(std_info$name[i], ": ",
format(mz(chr)[1], digits = 6), "-",
format(mz(chr)[2], digits = 6)),
peakCol = paste0(sample_colors[pks[, "column"]], 50),
peakBg = paste0(sample_colors[pks[, "column"]], 10))
## Indicate the selected peak position.
points(std_rt[i, ], std_max[i, ], col = sample_colors)
grid()
boxplot(log2(std_int), col = paste0(sample_colors, 80), las = 2,
names = chr$year_month, ylab = expression(log[2]~intensity))
points(x = seq_len(ncol(std_int)), log2(std_int[i, ]), pch = 16,
type = "o", cex = 1.5)
grid()
dev.off()
}
```
In addition to the standards we are also identifying chromatographic peaks for
the internal standards (code not shown).
```{r is-define, echo = FALSE}
is_info <- read.table(
"https://raw.githubusercontent.com/EuracBiomedicalResearch/lcms-standards/master/data/internal_standards.txt",
sep = "\t", header = TRUE, as.is = TRUE)
is_info <- is_info[!is.na(is_info[, "POS"]), ]
rownames(is_info) <- 1:nrow(is_info)
is_info$mzneut = NA
is_info$mz_ion = NA
for (i in seq(nrow(is_info))) {
if (grepl("C", is_info$formula[i])){
is_info$mzneut[i] <- getMolecule(
as.character(is_info$formula[i]))$exactmass
} else {
is_info$mzneut[i] = as.numeric(is_info$formula[i])
}
## Calculate also the m/z
is_info$mz_ion[i] <- unlist(
mass2mz(is_info$mzneut[i],
adduct = as.character(
is_info[i, "POS"])))
}
## Two-step approach: define first the rt/mz region for each standard and
## Extract the chromatogram afterwards.
is_info_mz_rt <- matrix(
ncol = 4, nrow = nrow(is_info),
dimnames = list(is_info$name, c("mzmin", "mzmax", "rtmin", "rtmax")))
for (i in 1:nrow(is_info)) {
pks <- chromPeaks(data_pos, mz = is_info$mz_ion[i], ppm = 50)
if (nrow(pks)) {
rtdiff <- abs(pks[, "rt"] - is_info$RT[i])
keep <- which(rtdiff < 40)
if (length(keep)) {
pks <- pks[keep, , drop = FALSE]
mzr <- range(pks[, c("mzmin", "mzmax")])
rtr <- range(pks[, c("rtmin", "rtmax")], is_info$RT[i])
is_info_mz_rt[i, ] <- c(mzr, rtr)
}
}
}
is_info <- is_info[!is.na(is_info_mz_rt[, "mzmin"]), ]
is_info_mz_rt <- is_info_mz_rt[!is.na(is_info_mz_rt[, "mzmin"]), ]
## Extract the ion chromatogram for each standard
rtr <- is_info_mz_rt[, c("rtmin", "rtmax")]
rtr[, 1] <- rtr[, 1] - 10
rtr[, 2] <- rtr[, 2] + 10
is_info_chr <- chromatogram(
data_pos, mz = is_info_mz_rt[, c("mzmin", "mzmax")],
rt = rtr, aggregationFun = "max")
## Plot the extracted ion chromatograms
dr <- paste0(IMAGE_PATH, "peakdetection/")
dir.create(dr, recursive = TRUE, showWarnings = FALSE)
sample_colors <- col_year_month[as.character(data_pos$year_month)]
for (i in 1:nrow(is_info)) {
cmpname <- is_info$name[i]
filename <- paste0(dr, "IS-", gsub("%", "p", cmpname, fixed = TRUE),
"_", "POS", ".png")
png(file = filename, width = 16, height = 8, units = "cm",
res = 300, pointsize = 4)
chr <- is_info_chr[i, ]
pks <- chromPeaks(chr)
plot(chr, col = "#00000040",
main = paste0(cmpname, ": ",
format(mz(chr)[1], digits = 6), "-",
format(mz(chr)[2], digits = 6)),
peakCol = paste0(sample_colors[pks[, "column"]], 50),
peakBg = paste0(sample_colors[pks[, "column"]], 10))
## This is just to indicate identified peaks with a rectangle.
xcms:::.add_chromatogram_peaks(
chr, pks,
col = paste0(sample_colors[pks[, "column"]], 40),
bg = NA,
type = "rectangle")
abline(v = is_info$RT[i])
dev.off()
}
is_info$rtmin <- NA_real_
is_info$rtmax <- NA_real_
## Add the min and max retention time for compounds we select for further
## analysis
is_info$rtmin[is_info$name == "Glycine (13C2, 99%; 15N, 99%)"] <- 165
is_info$rtmax[is_info$name == "Glycine (13C2, 99%; 15N, 99%)"] <- 175
is_info$rtmin[is_info$name == "L-Alanine (13C3, 99%; 15N, 99%)"] <- 165
is_info$rtmax[is_info$name == "L-Alanine (13C3, 99%; 15N, 99%)"] <- 175
is_info$rtmin[is_info$name == "L-Arginine HCl (13C6, 99%; 15N4, 99%)"] <- 170
is_info$rtmax[is_info$name == "L-Arginine HCl (13C6, 99%; 15N4, 99%)"] <- 200
is_info$rtmin[is_info$name == "L-Aspartic acid (13C4, 99%; 15N, 99%)"] <- 170
is_info$rtmax[is_info$name == "L-Aspartic acid (13C4, 99%; 15N, 99%)"] <- 195
is_info$rtmin[is_info$name == "L-Cystine (13C6, 99%; 15N2, 99%)"] <- 205
is_info$rtmax[is_info$name == "L-Cystine (13C6, 99%; 15N2, 99%)"] <- 230
is_info$rtmin[is_info$name == "L-Glutamic acid (13C5, 99%; 15N, 99%)"] <- 165
is_info$rtmax[is_info$name == "L-Glutamic acid (13C5, 99%; 15N, 99%)"] <- 190
is_info$rtmin[is_info$name == "L-Histidine HCl H2O (13C6; 15N3, 99%)"] <- 170
is_info$rtmax[is_info$name == "L-Histidine HCl H2O (13C6; 15N3, 99%)"] <- 210
is_info$rtmin[is_info$name == "L-Lysine 2HCl (13C6, 99%; 15N2, 99%)"] <- 170
is_info$rtmax[is_info$name == "L-Lysine 2HCl (13C6, 99%; 15N2, 99%)"] <- 210
is_info$rtmin[is_info$name == "L-Methionine (13C5, 99%; 15N, 99%)"] <- 155
is_info$rtmax[is_info$name == "L-Methionine (13C5, 99%; 15N, 99%)"] <- 167
is_info$rtmin[is_info$name == "L-Phenylalanine (13C9, 99%; 15N, 99%)"] <- 140
is_info$rtmax[is_info$name == "L-Phenylalanine (13C9, 99%; 15N, 99%)"] <- 163
is_info$rtmin[is_info$name == "L-Proline (13C5, 99%; 15N, 99%)"] <- 163
is_info$rtmax[is_info$name == "L-Proline (13C5, 99%; 15N, 99%)"] <- 177
is_info$rtmin[is_info$name == "L-Serine (13C3, 99%; 15N, 99%)"] <- 175
is_info$rtmax[is_info$name == "L-Serine (13C3, 99%; 15N, 99%)"] <- 190
is_info$rtmin[is_info$name == "L-Threonine (13C4, 99%; 15N, 99%)"] <- 165
is_info$rtmax[is_info$name == "L-Threonine (13C4, 99%; 15N, 99%)"] <- 185
is_info$rtmin[is_info$name == "L-Tyrosine (13C9, 99%; 15N, 99%)"] <- 165
is_info$rtmax[is_info$name == "L-Tyrosine (13C9, 99%; 15N, 99%)"] <- 180
is_info$rtmin[is_info$name == "L-Valine (13C5, 99%; 15N, 99%)"] <- 160
is_info$rtmax[is_info$name == "L-Valine (13C5, 99%; 15N, 99%)"] <- 172
is_info <- is_info[!is.na(is_info$rtmin), ]
## Re-create the m/z and rt range matrix
is_info_mz_rt <- matrix(
ncol = 4, nrow = nrow(is_info),
dimnames = list(is_info$name, c("mzmin", "mzmax", "rtmin", "rtmax")))
for (i in 1:nrow(is_info)) {
pks <- chromPeaks(data_pos, mz = is_info$mz_ion[i], ppm = 50,
rt = c(is_info$rtmin[i], is_info$rtmax[i]))
if (nrow(pks)) {
mzr <- range(pks[, c("mzmin", "mzmax")])
rtr <- range(pks[, c("rtmin", "rtmax")], is_info$RT[i])
rtr[1] <- rtr[1] - 1
rtr[2] <- rtr[2] + 1
is_info_mz_rt[i, ] <- c(mzr, rtr)
}
}
is_chr <- chromatogram(
data_pos, mz = is_info_mz_rt[, c("mzmin", "mzmax")],
rt = is_info_mz_rt[, c("rtmin", "rtmax")],
aggregationFun = "max")
fData(is_chr) <- cbind(fData(is_chr),
is_info[, c("name", "abbreviation", "formula",
"POS", "RT")])
## retention time and intensity matrices
is_rt <- matrix(ncol = ncol(is_chr), nrow = nrow(is_chr),
dimnames = dimnames(is_chr))
rownames(is_info) <- is_info$name
rownames(is_rt) <- rownames(is_info)
is_int <- is_rt
is_max <- is_int
for (i in 1:nrow(is_rt)) {
tmp <- is_chr[i, ]
tmp <- split.data.frame(chromPeaks(tmp), chromPeaks(tmp)[, "column"])
tmp <- lapply(tmp, function(z) z[which.max(z[, "maxo"]), ])
rts <- vapply(tmp, function(z) z["rt"], numeric(1))
ints <- vapply(tmp, function(z) z["into"], numeric(1))
maxs <- vapply(tmp, function(z) z["maxo"], numeric(1))
is_rt[i, names(rts)] <- rts
is_int[i, names(ints)] <- ints
is_max[i, names(maxs)] <- maxs
}
idx <- order(is_info$RT)
is_info <- is_info[idx, ]
is_chr <- is_chr[idx, ]
is_rt <- is_rt[idx, ]
is_int <- is_int[idx, ]
is_max <- is_max[idx, ]
```
The table of the (selected) internal standards is shown below.
```{r is-table, echo = FALSE, results = "asis"}
T <- data.frame(is_info[, c("name", "mz_ion")],
rt_mean = apply(is_rt, 1, mean, na.rm = TRUE),
rt_sd = apply(is_rt, 1, sd, na.rm = TRUE),
into_mean = apply(is_int, 1, mean, na.rm = TRUE),
into_cv = rowRsd(is_int, na.rm = TRUE))
pandoc.table(T, style = "rmarkdown",
caption = "Internal standards.")
```
With the exception of 3 standards, retention time shifts of all internal
standards are comparable. Also, the coefficient of variation are for most
between 0.4 and 0.5.
```{r export-eic-is, echo = FALSE, message = FALSE}
dr <- paste0(IMAGE_PATH, "internal-standards/")
dir.create(dr, recursive = TRUE, showWarnings = FALSE)
sample_colors <- col_year_month[as.character(data_pos$year_month)]
for (i in seq_len(nrow(is_chr))) {
fn <- paste0(dr, "IS-", gsub("%", "p", is_info$name[i], fixed = TRUE),
".png")
png(file = fn, width = 16, height = 10, units = "cm",
res = 300, pointsize = 4)
par(mfrow = c(1, 2), mar = c(5, 4.2, 2, 0.5))
chr <- is_chr[i, ]
pks <- chromPeaks(chr)
plot(chr, col = "#00000040",
main = paste0(is_info$name[i], ": ",
format(mz(chr)[1], digits = 6), "-",
format(mz(chr)[2], digits = 6)),
peakCol = paste0(sample_colors[pks[, "column"]], 50),
peakBg = paste0(sample_colors[pks[, "column"]], 10))
## Indicate the selected peak position.
points(is_rt[i, ], is_max[i, ], col = sample_colors)
grid()
boxplot(log2(is_int), col = paste0(sample_colors, 80), las = 2,
names = chr$year_month, ylab = expression(log[2]~intensity))
points(x = seq_len(ncol(is_int)), log2(is_int[i, ]), pch = 16,
type = "o", cex = 1.5)
grid()
dev.off()
}
```
## Comparison of batches
In this section we compare the signal of the selected features across the
batches. This complements the analysis performed on the base peak and total ion
signal performed above and aims at evaluating within and between-batch signals.
```{r batches-signal-boxplot, echo = FALSE, fig.width = 12, fig.height = 8, fig.cap = "Per sample distribution of base peak intensities (top) and signal distribution of selected ions (middle) and internal standards (bottom) across batches.", fig.path = IMAGE_PATH}
par(mfrow = c(3, 1), mar = c(1, 4.5, 1, 1))
boxplot(lapply(bpchr, function(z) log2(intensity(z))), outline = FALSE,
ylab = expression(log[2]~base~peak~intensity), xaxt = "n",
main = "QC samples, positive polarity",
col = paste0(col_year_month[bpchr$year_month], 40))
grid(nx = NA, ny = NULL)
boxplot(log2(std_int), ylab = expression(log[2]~intensity), xaxt = "n",
main = "Selected ions",
col = paste0(col_year_month[bpchr$year_month], 40))
grid(nx = NA, ny = NULL)
boxplot(log2(is_int), ylab = expression(log[2]~intensity), xaxt = "n",
main = "Internal standards",
col = paste0(col_year_month[bpchr$year_month], 40))
grid(nx = NA, ny = NULL)
```
Considering only the selected ions or the internal satndards, differences within
each batch and also between batches (except from the 2 first samples) are
smaller than for the base peak signal. This suggests the base peak signal to be
eventually more affected by background or non-specific noise. Also, differences
in average signal for the last 5 samples is very small suggesting that the data
within this batch (and the previous one) is (relatively) reproducible.
We next evaluate the signal across samples separately for each feature.
```{r batches-signal-per-feature, echo = FALSE, fig.width = 8, fig.height = 4, fig.cap = "Signal of selected ions across batches/samples.", fig.path = IMAGE_PATH}
par(mfrow = c(1, 1), mar = c(1, 4.5, 1, 1))
plot(xlim = c(1, ncol(std_int)), ylim = range(log2(std_int), na.rm = TRUE),
3, 3, pch = NA, ylab = expression(log[2]~intensity),
main = "Selected ions")
grid()
for (i in seq_len(nrow(std_int))) {
points(x = seq_len(ncol(std_int)), log2(std_int[i, ]), col = "#00000060",
type = "l")
}
```
Most of the features show the same pattern across samples. This suggests
*global* batch-specific signal shifts which could be adjusted by the
between-batch normalization. There are however some features which have
considerably lower intensities in the last batches from September 2020 (one
example being Glycero-phosphocholine which was not detected in most of these
batches).
To further evaluate whether the differences in feature concentrations are
related to global signal differences we plot the total ion signal of each sample
against the average intensity of the standards in that sample.
```{r batches-ti-vs-avg-int, echo = FALSE, fig.width = 12, fig.height = 6, fig.cap = "Total ion signal (x-axis) against average intensity of selected standards (left) and internal standards (right) per sample.", fig.path = IMAGE_PATH}
ti <- log2(vapply(tichr, function(z) sum(intensity(z), na.rm = TRUE),
numeric(1)))
std_median <- apply(log2(std_int), 2, median, na.rm = TRUE)
is_median <- apply(log2(is_int), 2, median, na.rm = TRUE)
par(mfrow = c(1, 2))
plot(ti, std_median,
xlab = "total ion intensity", ylab = "median intensity of standards",
main = "Selected ions")
L <- lm(std_median ~ ti)
abline(L)
plot(ti, is_median,
xlab = "total ion intensity", ylab = "median intensity of standards",
main = "Internal standards")
L <- lm(is_median ~ ti)
abline(L)
```
Indeed, the average intensity of peaks of the selected standards and for the
internal standards is related to the total signal measured by the MS. Next we
evaluate for each standard its relationship with the total ion signal.
```{r std-int-to-ti-table, echo = FALSE, results = "asis"}
std_lm <- vector("list", nrow(std_int))
for (i in seq_len(nrow(std_int))) {
y <- log2(std_int[i, ])
std_lm[[i]] <- lm(y ~ ti)
}
T <- data.frame(name = rownames(std_int),
slope = vapply(std_lm, function(z) z$coefficients[2],
numeric(1)),
r.squared = vapply(std_lm, function(z) summary(z)$r.squared,
numeric(1)))
pandoc.table(
T, style = "rmarkdown",
caption = paste0("Relationship between intensities of standards to the ",
"total ion signal of a sample."))
```
For many standards there is a correlation between the intensities and the total
ion signal of the samples. Some however, specifically *Sphingosine*, *Inosine*,
*3-Nitrotyrosine*, *Glycine*, *ADMA*, *Lysine*, *Glycero-phosphocholine* and
*Cystine* this correlation is either absent or poor. We will investigate this in
more detail in the next section.
Next we perform the same analysis for the internal standards.
```{r is-int-to-ti-table, echo = FALSE, results = "asis"}
is_lm <- vector("list", nrow(is_int))
for (i in seq_len(nrow(is_int))) {
y <- log2(is_int[i, ])
is_lm[[i]] <- lm(y ~ ti)
}
T <- data.frame(name = rownames(is_int),
slope = vapply(is_lm, function(z) z$coefficients[2],
numeric(1)),
r.squared = vapply(is_lm, function(z) summary(z)$r.squared,
numeric(1)))
pandoc.table(
T, style = "rmarkdown",
caption = paste0("Relationship between intensities of internal standards ",
"to the total ion signal of a sample."))
```
For most internal standards their intensity correlates with the total ion signal
per sample. Exceptions are *Glycine* and *L-Cystine*.
## Feature-wise analysis
In this analysis we evaluate whether some features are more affected by
intensity differences or retention time shifts than others and try to find
possible explanations for this. At first we evaluate whether some features are
more affected by retention time shifts (and/or intensity differences) than
others. Thus we evaluate if the retention time shifts are global or
feature-specific (or rt) dependent.
```{r feature-rt-shift-boxplot, fig.path = IMAGE_PATH, fig.width = 7, fig.height = 7, fig.cap = "Per-feature retention time difference across the analyzed batches."}
idx_rt <- order(std_info$rtmin)
par(mar = c(7, 4.2, 1, 1))
boxplot(t(std_rt[idx_rt, ]), las = 2, ylab = "retention time")
grid()
```
Some standards show only very small variations in retention time while others
have large retention time shifts between batches. It seems that compounds
eluting very early are less affected by retention time shifts (because they in
fact are not retained in the column). Interestingly, most standards with large
retention time shifts elute at around 180 seconds. However, at that time there
are also standards that are not or much less affected by retention time shifts.
Next we ask the question:
- is the retention time shift dependent on the feature's m/z or intensity?
To evaluate this potential dependency we plot the standard deviation of the
feature's retention time against its m/z or average intensity (across all
samples).
```{r feature-rt-shift-dependency, fig.path = IMAGE_PATH, fig.width = 15, fig.height = 5, fig.cap = "Per-feature retention time difference across the analyzed batches."}
par(mfrow = c(1, 3))
plot(std_info$mz_ion, apply(std_rt, 1, sd, na.rm = TRUE),
xlab = "m/z", ylab = "sd retention time")
grid()
plot(apply(log2(std_int), 1, mean, na.rm = TRUE),
apply(std_rt, 1, sd, na.rm = TRUE),
xlab = "mean log2 intensity", ylab = "sd retention time")
grid()
plot(apply(log2(std_int), 1, sd, na.rm = TRUE),
apply(std_rt, 1, sd, na.rm = TRUE),
xlab = "sd log2 intensity", ylab = "sd retention time")
grid()
```
There seems to be no dependency between m/z or intensity values and retention
time shifts and also no relationship between variation in retention time and
variation in intensity.
Next evaluate the per-feature variance in intensities (across samples).
```{r feature-int-boxplot, fig.path = IMAGE_PATH, fig.width = 9, fig.height = 7, fig.cap = "Per-feature intensity difference across the analyzed batches."}
par(mar = c(9, 4.2, 1, 1))
boxplot(log2(t(std_int[idx_rt, ])), las = 2, ylab = "log2 intensity")
grid(ny = NULL, nx = NA)
```
Intensity differences seem to be feature-specific, with high abundance features
seeming to have lower variance. To evaluate a potential relationship between
intensity and variance we plot below the average intensity against its standard
deviation for each feature.
```{r feature-int-mean-vs-sd, fig.path = IMAGE_PATH, fig.width = 7, fig.height = 7, fig.cap = "Average intensity for each feature against its standard deviation."}
plot(apply(log2(std_int), 1, sd, na.rm = TRUE),
apply(log2(std_int), 1, mean, na.rm = TRUE),
xlab = "sd log2 intensity", ylab = "mean log2 intensity")
grid(ny = NULL, nx = NA)
```
Again, no clear dependency is detectable. Thus, the variance in intensity seems
to be feature specific, but not dependent on the total signal of the standard.
As we have already seen in the previous section, for most standards the
intensity follows the total ion signal of the sample (and hence batch). Only for
a handful of features there was no such relationship. We are next trying to
understand why these standards yield different intensities between batches. At
first we evaluate the full scan at peak position for these standards in each
sample. We thus extract next the full scan for each standard.
```{r extract-full-scan}
## Replace NAs with median of row.
std_rt_nona <- std_rt
for (i in seq_len(nrow(std_rt))) {
nas <- is.na(std_rt[i, ])
if (any(nas))
std_rt_nona[i, nas] <- median(std_rt[i, !nas])
}
## We want to get for each sample the spectrum at the peak position, or if no
## peak was identified on the median rt in all other samples.
## As a result we want to get a list (each element being for one standard) of
## Spectra objects (one spectrum per sample/file).
which_rt_file <- function(rt, file, x) {
sel <- fromFile(x) == file
idx <- which.min(abs(rtime(x)[sel] - rt))
which.max(sel) + idx - 1
}
std_sp <- vector("list", nrow(std_info))
names(std_sp) <- rownames(std_info)
for (i in seq_along(std_sp)) {
idx <- mapply(which_rt_file, std_rt_nona[i, ],
seq_len(ncol(std_rt_nona)), MoreArgs = list(x = data_pos))
std_sp[[i]] <- Spectra(extractSpectraData(data_pos[idx]))
}
```
The first standard is *Sphingosine*. As we can see in the plot below, the
intensities measured in September 2020 are much lower (compared to the median
intensity of all standards) than in all other batches.
We next calculate m/z values for all potential adducts of the compound and
filter the full scan to only contain these.
```{r}
i <- which(std_info$name == "Sphingosine")
sps <- std_sp[[i]]
adds <- mass2mz(std_info$mzneut[i],
adduct = adducts(polarity = "positive"))[[1]]
adds <- sort(adds[adds > 0 & adds < 1000])
sps_adds <- filterMzValues(sps, mz = adds, ppm = 20)
```
Next we compare the MS1 spectrum containing only peaks representing potential
adducts of the compound (*ion spectrum*) for Sphingosine in a sample from
February 2020 to one from September 2020.
```{r sphingosine-ion-spectrum-mirror-14-15, echo = FALSE, fig.path = IMAGE_PATH, fig.width = 8, fig.height = 8}
label_function <- function(x) {
names(adds)[MsCoreUtils::closest(mz(x)[[1L]], adds,
tolerance = 0, ppm = 40,
duplicates = "closest")]
}
plotSpectraMirror(sps_adds[14], sps_adds[15], ppm = 40,
main = paste0(std_info$name[i], "; sample 14/15"),
labels = label_function, labelPos = 4, labelSrt = 45)