autonomics: platform-aware analysis
read + preprocess + analyze + plot cross-platform omics data
5 June 2024
Contents
Abstract
autonomics is created to make cross-platform omics analysis intuitive and enjoyable. It contains import + preprocess + analyze + visualize one-liners for RNAseq, MS proteomics, SOMAscan and Metabolon platforms and a generic importer for data from any rectangular omics file. With a focus on not only automation but also customization, these importers have a flexible formula/ block/contrastdefs interface which allows to define any statistical formula, fit any general linear model, quantify any contrast, and use random effects or precision weights if required, building on top of the powerful limma platform for statistical analysis. It also offers exquisite support for analyzing complex designs such as the innovative contrastogram visualization to unravel and summarize the differential expression structure in complex designs. By autonomating repetitive tasks, generifying common steps, and intuifying complex designs, it makes cross-platform omics data analysis a fun experience. Try it and enjoy :-).
Autonomics offers import + preprocess + analyze + visualize one-liners for RNAseq, MS proteomics, SOMAscan and Metabolon platforms, as well a generic importer for data from any rectangular omics file. We will discuss each of these in more detail in the following sections, but all of them follow the following general steps:
- read
- exprs, features, samples
- extract subgroups from sampleids / samplefile
- wrap these into a SummarizedExperiment
- preprocess
- normalize
- log2 transform
- impute in a platform-aware fashion
- filter
- analyze
- plot sample distributions
- plot detections (absent/present)
- plot pca biplots
- fit the linear model
~0+subgroup - quantify contrasts between subsequent factor levels and plot volcanoes
1 RNAseq
Autonomics provides ready-to-use importers for both count as well as BAM files, which read / preprocess / analyze RNAseq data. Specific to RNAseq is counting reads using Rsubread::featureCounts (for read_rna_bams)), normalizing for library composition (cpm) or gene length (tpm), estimating voom precision weights, and using these in linear modeling. Let’s illustrate both of these on example datasets:
read_rnaseq_counts
Billing et al. (2019) studied the differentiation of embryonic (E00) into mesenchymal stem cells (E00.8, E01, E02, E05, E15, E30), with similar properties as bone-marrow mesenchymal stem cells (M00). Importing the RNAseq counts:
require(autonomics, quietly = TRUE)
##
## Attaching package: 'autonomics'
## The following objects are masked from 'package:stats':
##
## biplot, loadingsfile <- system.file('extdata/billing19.rnacounts.txt', package = 'autonomics')
object <- read_rnaseq_counts(file = file, fit = 'limma', plot = TRUE, label = 'gene')
## Read /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/billing19.rnacounts.txt
## Infer subgroup from sample_ids
## Infer subgroup from sample_ids
## Preprocess
## Keep 24/24 features: count >= 10 (~subgroup)
## counts: add 0.5
## cpm: tmm scale libsizes
## cpm
## voom: ~subgroup
## counts: rm 0.5
## Add PCA
## LinMod
## Code subgroup
## level
## coefficient E00 E00.8 E01 E02 E05 E15 E30 M00
## Intercept 1 . . . . . . .
## E00.8 -1 1 . . . . . .
## E01 -1 . 1 . . . . .
## E02 -1 . . 1 . . . .
## E05 -1 . . . 1 . . .
## E15 -1 . . . . 1 . .
## E30 -1 . . . . . 1 .
## M00 -1 . . . . . . 1
## lmFit(~subgroup, weights = assays(object)$weights)
## coefficient fit downfdr upfdr downp upp
## <fctr> <fctr> <int> <int> <int> <int>
## 1: Intercept limma 0 24 0 24
## 2: E00.8 limma 1 2 1 4
## 3: E01 limma 3 2 5 3
## 4: E02 limma 2 2 3 4
## 5: E05 limma 1 2 2 2
## 6: E15 limma 8 9 8 10
## 7: E30 limma 7 10 8 10
## 8: M00 limma 6 7 6 7
## Plot summary
read_rnaseq_counts(sfile)
The sfile/sfileby arguments in both importers allow to merge in sample data from a separate file by a specified column. This is especially useful in clinical datasets, such as the GSE161731 Covid-19 dataset, for which both a counts file as well as a sample file can be downloaded from GEO:
basedir <- file.path(tempdir(), 'datasets')
dir.create(basedir, showWarnings = FALSE, recursive = TRUE)
if (!dir.exists(file.path(basedir, 'GSE161731'))){
sink <- GEOquery::getGEOSuppFiles("GSE161731", baseDir = basedir)
}
## Setting options('download.file.method.GEOquery'='auto')
## Setting options('GEOquery.inmemory.gpl'=FALSE)object <- read_rnaseq_counts(
file = file.path(basedir, 'GSE161731/GSE161731_counts.csv.gz'),
sfile = file.path(basedir, 'GSE161731/GSE161731_counts_key.csv.gz'),
by.y = 'rna_id',
formula = ~ gender,
plot = TRUE)
## Read /tmp/RtmpsgvrVb/datasets/GSE161731/GSE161731_counts.csv.gz
## Merge sdata: /tmp/RtmpsgvrVb/datasets/GSE161731/GSE161731_counts_key.csv.gz
## Preprocess
## Keep 18974/60675 features: count >= 10 (~gender)
## counts: add 0.5
## cpm: tmm scale libsizes
## cpm
## voom: ~gender
## counts: rm 0.5
## Add PCA
## LinMod
## Code gender
## level
## coefficient Female Male
## Intercept 1 .
## Male -1 1
## lmFit(~gender, weights = assays(object)$weights)
## coefficient fit downfdr upfdr downp upp
## <fctr> <fctr> <int> <int> <int> <int>
## 1: Intercept limma 2990 14603 3005 14613
## 2: Male limma 40 74 1207 1447
## Plot summary
## Warning: ggrepel: 96 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps
read_rnaseq_counts(block)
The block argument in both importers allows to model e.g. subject_id as a random effect, which is very common in clinical data. Repeating the GEO Covid-19 analysis with subject_id as a block effect. Outcomment the line below, which takes a few minutes to complete.
#object <- read_rnaseq_counts(
# file = '~/autonomicscache/datasets/GSE161731/GSE161731_counts.csv.gz',
# sfile = '~/autonomicscache/datasets/GSE161731/GSE161731_counts_key.csv.gz',
# sfileby = 'rna_id',
# subgroupvar = 'gender',
# block = 'subject_id')read_rnaseq_bams
Billing et al. (2016) compared three types of stem cells: embryonic (E), embryonic differentiated into mesenchymal (EM), and bone-marrow mesenchymal (BM). Importing a downsized version of the RNAseq BAM files (with only a subset of all reads):
# not run to avoid issues with R CMD CHECK
if (requireNamespace('Rsubread')){
object <- read_rnaseq_bams(
dir = download_data('billing16.bam.zip'),
paired = TRUE,
genome = 'hg38',
pca = TRUE,
fit = 'limma',
plot = TRUE)
}cpm/tmm/voom effects on power
Proper preprocessing leads to increased power:
file <- system.file('extdata/billing19.rnacounts.txt', package = 'autonomics')
# log2counts
object <- read_rnaseq_counts(file,
cpm = FALSE, voom = FALSE, fit = 'limma', verbose = FALSE, plot = FALSE)
## LinMod
## Code subgroup
## level
## coefficient E00 E00.8 E01 E02 E05 E15 E30 M00
## Intercept 1 . . . . . . .
## E00.8 -1 1 . . . . . .
## E01 -1 . 1 . . . . .
## E02 -1 . . 1 . . . .
## E05 -1 . . . 1 . . .
## E15 -1 . . . . 1 . .
## E30 -1 . . . . . 1 .
## M00 -1 . . . . . . 1 colSums(summarize_fit(fdt(object), 'limma')[-1, c(3,4)])
## downfdr upfdr
## 7 13# log2cpm
object <- read_rnaseq_counts(file,
cpm = TRUE, voom = FALSE, fit = 'limma', verbose = FALSE, plot = FALSE)
## LinMod
## Code subgroup
## level
## coefficient E00 E00.8 E01 E02 E05 E15 E30 M00
## Intercept 1 . . . . . . .
## E00.8 -1 1 . . . . . .
## E01 -1 . 1 . . . . .
## E02 -1 . . 1 . . . .
## E05 -1 . . . 1 . . .
## E15 -1 . . . . 1 . .
## E30 -1 . . . . . 1 .
## M00 -1 . . . . . . 1 colSums(summarize_fit(fdt(object), 'limma')[-1, c(3,4)])
## downfdr upfdr
## 29 34# log2cpm + voom
object <- read_rnaseq_counts(file, # log2 cpm + voom
cpm = TRUE, voom = TRUE, fit = 'limma', verbose = FALSE, plot = FALSE)
## LinMod
## Code subgroup
## level
## coefficient E00 E00.8 E01 E02 E05 E15 E30 M00
## Intercept 1 . . . . . . .
## E00.8 -1 1 . . . . . .
## E01 -1 . 1 . . . . .
## E02 -1 . . 1 . . . .
## E05 -1 . . . 1 . . .
## E15 -1 . . . . 1 . .
## E30 -1 . . . . . 1 .
## M00 -1 . . . . . . 1 colSums(summarize_fit(fdt(object), 'limma')[-1, c(3,4)])
## downfdr upfdr
## 28 342 Proteingroups/phosphosites
A popular approach in (DDA) MS proteomics data analysis is to use MaxQuant to produce proteinGroups.txt and phospho (STY)Sites.txt files. These can be imported using read_proteingroups / read_phosphosites, which :
- Read
- quantification cols: LFQ intensities, Normalized Ratios, or Reporter intensities
- feature cols: id, Majority protein IDs, Protein names, Gene names, Contaminant, Potential contaminant, Reverse
- subgroups from sampleids / samplefile
- Preprocess
- Convert 0 -> NA (they represent lack of detection)
- Log2 transform
- Rm contaminants/reverse/unreplicated features
- Impute consistent NAs with values drawn from a half-normal distribution around zero
- For phosphosites: compute occupancies by subtracting protein expressions
- Analyze
- plot sample distributions
- plot detections
- pca biplot
- Fit a linear model
- Quantify contrasts between subsequent factor levels and plot volcanoes
LFQ intensities
An LFQ intensities example is the Fukuda et al. (2020) dataset, which compares the proteome of 30d zebrafish juveniles versus adults using label-free quantification. In the volcano plot measured values are shown with circles, imputed values as triangles.
file <- system.file('extdata/fukuda20.proteingroups.txt', package = 'autonomics')
object <- read_maxquant_proteingroups(file = file, plot = TRUE)
## Read proteingroups /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/fukuda20.proteingroups.txt
## contaminanthdrs /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/contaminants.tsv
## maxquanthdrs
## Annotate Uncollapse 20 proIds into 32 proteins
## Drop REV_ 32 proteins
## Annotate 0 using uniprothdrs
## 1 using contaminanthdrs
## 29 using maxquanthdrs
## 0 using uniprotrestapi
## Filter 32 proteins
## within proId 31 proteins swissprot > trembl
## 25 proteins fullseq > fragment
## 20 proteins protein > transcript > homolog > prediction > uncertain
## Add REV__ 20 proteins
## Recollapse 20 proIds
## SumExp
## Replace 0->NA for 27/120 values (in 11/20 features of 6/6 samples)
## Standardize snames: LFQ intensity 30dpt.R1 -> 30dpt.R1
## Infer subgroup from sample_ids
## Retain 19/20 features: ~reverse == ""
## Retain 18/19 features: ~contaminant == ""
## Add PCA
## LinMod
## Code subgroup
## level
## coefficient X30dpt Adult
## Intercept 1 .
## Adult -1 1
## lmFit(~subgroup)
## coefficient fit downfdr upfdr downp upp
## <fctr> <fctr> <int> <int> <int> <int>
## 1: Intercept limma 0 18 0 18
## 2: Adult limma 2 2 3 2
## Plot summary
Normalized ratios
Normalized ratios were used in the proteomics profiling of the earlier described Billing et al. (2019) study, which examined the differentiation of embryonic stem cells (E00) into mesenchymal stem cells (E01,E02, E05, E15, E30), and compared these to bone-marrow derived mesenchymal stem cells (M00). The proteomics profiling experiment was performed using MS1 labeling: a light (L) labeled internal standard was created by mixing all samples in equal amounts, the subsequent samples were either medium (M) or heavy (H) labeled. Not all label combinations were of interest, and the select_subgroups argument allows to specifies which subgroup combinations to retain:
file <- system.file('extdata/billing19.proteingroups.txt', package = 'autonomics')
subgroups <- c('E00_STD', 'E01_STD', 'E02_STD', 'E05_STD', 'E15_STD', 'E30_STD', 'M00_STD')
object <- read_maxquant_proteingroups(file = file, subgroups = subgroups, plot = TRUE)
## Read proteingroups /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/billing19.proteingroups.txt
## contaminanthdrs /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/contaminants.tsv
## maxquanthdrs
## Annotate Uncollapse 25 proIds into 83 proteins
## Drop REV_ 45 proteins
## Annotate 0 using uniprothdrs
## 1 using contaminanthdrs
## 34 using maxquanthdrs
## 0 using uniprotrestapi
## Filter 45 proteins
## within proId 24 proteins swissprot > trembl
## Add REV__ 24 proteins
## Recollapse 25 proIds
## SumExp
## Replace NaN->NA for 175/825 values (in 12/25 features of 33/33 samples)
## Standardize snames: Ratio M/L normalized STD(L).E00(M).E01(H).R1 -> STD(L).E00(M).E01(H).R1{M/L}
## Infer subgroup from sample_ids
## Retain 21/33 samples: ~subgroup %in% subgroups
## Retain 24/25 features: ~reverse == ""
## Retain 23/24 features: ~contaminant == ""
## Retain 22/23 features: non-zero, non-NA, and non-NaN for some sample
## Add PCA
## LinMod
## Code subgroup
## level
## coefficient E00_STD E01_STD E02_STD E05_STD E15_STD E30_STD M00_STD
## Intercept 1 . . . . . .
## E01_STD -1 1 . . . . .
## E02_STD -1 . 1 . . . .
## E05_STD -1 . . 1 . . .
## E15_STD -1 . . . 1 . .
## E30_STD -1 . . . . 1 .
## M00_STD -1 . . . . . 1
## lmFit(~subgroup)
## coefficient fit downfdr upfdr downp upp
## <fctr> <fctr> <int> <int> <int> <int>
## 1: Intercept limma 1 5 2 6
## 2: E01_STD limma 0 0 0 1
## 3: E02_STD limma 0 2 1 2
## 4: E05_STD limma 0 0 0 3
## 5: E15_STD limma 2 8 2 11
## 6: E30_STD limma 2 8 2 8
## 7: M00_STD limma 2 5 2 6
## Plot summary
The phosphosites can be read in a similar fashion:
fosfile <- system.file('extdata/billing19.phosphosites.txt', package = 'autonomics')
profile <- system.file('extdata/billing19.proteingroups.txt', package = 'autonomics')
subgroups <- c('E00_STD', 'E01_STD', 'E02_STD', 'E05_STD', 'E15_STD', 'E30_STD', 'M00_STD')
object <- read_maxquant_phosphosites(fosfile = fosfile, profile = profile, subgroups = subgroups, plot = TRUE)
## Read proteingroups /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/billing19.proteingroups.txt
## phosphosites /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/billing19.phosphosites.txt
## Read 21 phosphosites in proteins, contaminants, reverse
## Retain 21 phosphosites in single proteingroup
## Retain 21 phosphosites with signal in some sample
## Keep proteingroup uniprots
## contaminanthdrs /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/contaminants.tsv
## maxquanthdrs
## Annotate Uncollapse 21 fosIds into 53 proteins
## Drop REV_ 28 proteins
## Annotate 0 using uniprothdrs
## 1 using contaminanthdrs
## 20 using maxquanthdrs
## 0 using uniprotrestapi
## Filter 28 proteins
## within fosId 16 proteins swissprot > trembl
## Add REV__ 16 proteins
## Recollapse 21 fosIds
## SumExp
## Replace NaN->NA for 104/693 values (in 6/21 features of 33/33 samples)
## Infer subgroup from sample_ids
## Retain 21/33 samples: ~subgroup %in% subgroups
## Retain 20/21 features: ~reverse == ""
## Retain 19/20 features: ~contaminant == ""
## Add PCA
## LinMod
## Code subgroup
## level
## coefficient E00_STD E01_STD E02_STD E05_STD E15_STD E30_STD M00_STD
## Intercept 1 . . . . . .
## E01_STD -1 1 . . . . .
## E02_STD -1 . 1 . . . .
## E05_STD -1 . . 1 . . .
## E15_STD -1 . . . 1 . .
## E30_STD -1 . . . . 1 .
## M00_STD -1 . . . . . 1
## lmFit(~subgroup)
## coefficient fit downfdr upfdr downp upp
## <fctr> <fctr> <int> <int> <int> <int>
## 1: Intercept limma 3 3 5 3
## 2: E01_STD limma 1 1 1 2
## 3: E02_STD limma 2 0 3 2
## 4: E05_STD limma 1 1 1 4
## 5: E15_STD limma 1 6 2 6
## 6: E30_STD limma 0 4 1 5
## 7: M00_STD limma 2 3 2 4
## Plot summary
3 Metabolon
Metabolon delivers metabolomics results in the form of an xlsx file, with values of interest likely in the (second) OrigScale sheet. Features are laid out in rows, samples in columns, and additional feature/sample data are self-contained in the file. Metabolon data can be read/analyzed with the autonomics function read_metabolon, as illustrated below on the dataset of Halama and and Atkin (2018), who investigates the effects of hypoglycemia on a number of subjects (SUB) at four different time points (t0, t1, t2, t3):
file <- system.file('extdata/atkin.metabolon.xlsx', package = 'autonomics')
object <- read_metabolon(file, subgroupvar = 'subgroup', block = 'Subject', plot = TRUE)
## Read: /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/atkin.metabolon.xlsx
## Infer subgroup from sample_ids
## log2 metabolon
## Add PCA
## LinMod
## Code subgroup
## level
## coefficient t0 t1 t2 t3
## Intercept 1 . . .
## t1 -1 1 . .
## t2 -1 . 1 .
## t3 -1 . . 1
## Dupcor `Subject`
## lmFit(~subgroup | Subject)
## coefficient fit downfdr upfdr downp upp
## <fctr> <fctr> <int> <int> <int> <int>
## 1: Intercept limma 0 20 0 20
## 2: t1 limma 8 1 10 1
## 3: t2 limma 7 3 8 3
## 4: t3 limma 1 1 4 2
## Plot summary
4 Somascan
Somascan results are returned in a txt file with extension .adat. Features are laid out in columns and samples in rows. Rectangles with feature/sample data are also contained in the file. The autonomics function read_somascan reads/analyzes SOMAscan files, additionally filtering samples/features for quality and type. This is illustrated on the above described dataset from Halama and and Atkin (2018), who investigated the effects of hypoglycemia on a number of subjects (SUB) at four different time points (t0, t1, t2, t3):
file <- system.file('extdata/atkin.somascan.adat', package = 'autonomics')
object <- read_somascan(file, block = 'Subject', plot = TRUE)
## Read: /tmp/RtmphdRJQu/Rinst171f806f813fd2/autonomics/extdata/atkin.somascan.adat
## log2 somascan
## Add PCA
## LinMod
## Code subgroup
## level
## coefficient t0 t1 t2 t3
## Intercept 1 . . .
## t1 -1 1 . .
## t2 -1 . 1 .
## t3 -1 . . 1
## Dupcor `Subject`
## lmFit(~subgroup | Subject)
## coefficient fit downfdr upfdr downp upp
## <fctr> <fctr> <int> <int> <int> <int>
## 1: Intercept limma 0 20 0 20
## 2: t1 limma 0 1 1 3
## 3: t2 limma 2 2 2 4
## 4: t3 limma 0 0 0 0
## Plot summary
SessionInfo
sessionInfo()
## R version 4.4.0 (2024-04-24)
## Platform: x86_64-pc-linux-gnu
## Running under: Ubuntu 22.04.4 LTS
##
## Matrix products: default
## BLAS: /home/biocbuild/bbs-3.19-bioc/R/lib/libRblas.so
## LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.10.0
##
## locale:
## [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
## [3] LC_TIME=en_GB LC_COLLATE=C
## [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
## [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
## [9] LC_ADDRESS=C LC_TELEPHONE=C
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
##
## time zone: America/New_York
## tzcode source: system (glibc)
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] autonomics_1.12.1 BiocStyle_2.32.0
##
## loaded via a namespace (and not attached):
## [1] DBI_1.2.3 gridExtra_2.3
## [3] readxl_1.4.3 rlang_1.1.4
## [5] magrittr_2.0.3 matrixStats_1.3.0
## [7] compiler_4.4.0 RSQLite_2.3.7
## [9] reshape2_1.4.4 vctrs_0.6.5
## [11] stringr_1.5.1 pkgconfig_2.0.3
## [13] crayon_1.5.2 fastmap_1.2.0
## [15] magick_2.8.3 dbplyr_2.5.0
## [17] XVector_0.44.0 labeling_0.4.3
## [19] utf8_1.2.4 rmarkdown_2.27
## [21] tzdb_0.4.0 UCSC.utils_1.0.0
## [23] tinytex_0.51 purrr_1.0.2
## [25] bit_4.0.5 xfun_0.44
## [27] MultiAssayExperiment_1.30.2 zlibbioc_1.50.0
## [29] cachem_1.1.0 GenomeInfoDb_1.40.1
## [31] jsonlite_1.8.8 fractional_0.1.3
## [33] blob_1.2.4 highr_0.11
## [35] DelayedArray_0.30.1 BiocParallel_1.38.0
## [37] tweenr_2.0.3 codingMatrices_0.4.0
## [39] parallel_4.4.0 R6_2.5.1
## [41] bslib_0.7.0 stringi_1.8.4
## [43] RColorBrewer_1.1-3 limma_3.60.2
## [45] GenomicRanges_1.56.0 jquerylib_0.1.4
## [47] cellranger_1.1.0 Rcpp_1.0.12
## [49] bookdown_0.39 SummarizedExperiment_1.34.0
## [51] knitr_1.47 R.utils_2.12.3
## [53] readr_2.1.5 IRanges_2.38.0
## [55] Matrix_1.7-0 igraph_2.0.3
## [57] tidyselect_1.2.1 abind_1.4-5
## [59] yaml_2.3.8 codetools_0.2-20
## [61] curl_5.2.1 plyr_1.8.9
## [63] lattice_0.22-6 tibble_3.2.1
## [65] rARPACK_0.11-0 Biobase_2.64.0
## [67] withr_3.0.0 evaluate_0.23
## [69] polyclip_1.10-6 BiocFileCache_2.12.0
## [71] xml2_1.3.6 pillar_1.9.0
## [73] BiocManager_1.30.23 filelock_1.0.3
## [75] MatrixGenerics_1.16.0 stats4_4.4.0
## [77] ellipse_0.5.0 generics_0.1.3
## [79] hms_1.1.3 S4Vectors_0.42.0
## [81] ggplot2_3.5.1 munsell_0.5.1
## [83] scales_1.3.0 glue_1.7.0
## [85] tools_4.4.0 data.table_1.15.4
## [87] RSpectra_0.16-1 locfit_1.5-9.9
## [89] GEOquery_2.72.0 grid_4.4.0
## [91] tidyr_1.3.1 edgeR_4.2.0
## [93] colorspace_2.1-0 GenomeInfoDbData_1.2.12
## [95] ggforce_0.4.2 cli_3.6.2
## [97] fansi_1.0.6 mixOmics_6.28.0
## [99] S4Arrays_1.4.1 rematch_2.0.0
## [101] dplyr_1.1.4 corpcor_1.6.10
## [103] pcaMethods_1.96.0 gtable_0.3.5
## [105] R.methodsS3_1.8.2 sass_0.4.9
## [107] digest_0.6.35 BiocGenerics_0.50.0
## [109] SparseArray_1.4.8 ggrepel_0.9.5
## [111] farver_2.1.2 memoise_2.0.1
## [113] htmltools_0.5.8.1 R.oo_1.26.0
## [115] lifecycle_1.0.4 httr_1.4.7
## [117] statmod_1.5.0 bit64_4.0.5
## [119] MASS_7.3-60.2References
Appendix
A M Billing, S S Dib, A M Bhagwat, I T da Silva, R D Drummond, S Hayat, R Al-Mismar, H Ben-Hamidane, N Goswami, K Engholm-Keller, M R Larsen, K Suhre, A Rafii, J Graummann (2019). Mol Cell Proteomics. 18(10):1950-1966. doi: 10.1074/mcp.RA119.001356.
A M Billing, H Ben Hamidane, S S Dib, R J Cotton, A M Bhagwat, P Kumar, S Hayat, N A Yousri, N Goswami, K Suhre, A Rafii, J Graumann (2016). Comprehensive transcriptomics and proteomics characterization of human mesenchymal stem cells reveals source specific cellular markers. Sci Rep. 9;6:21507. doi: 10.1038/srep21507.
R Fukuda, R Marin-Juez, H El-Sammak, A Beisaw, R Ramadass, C Kuenne, S Guenther, A Konzer, A M Bhagwat, J Graumann, D YR Stainier (2020). EMBO Rep. 21(8): e49752. doi: 10.15252/embr.201949752
A Halama, M Kulinski, S Dib, S B Zaghlool, K S Siveen, A Iskandarani, J Zierer 3, K S Prabhu, N J Satheesh, A M Bhagwat, S Uddin, G Kastenmüller, O Elemento, S S Gross, K Suhre (2018). Accelerated lipid catabolism and autophagy are cancer survival mechanisms under inhibited glutaminolysis. Cancer Lett., 430:133-147. doi:10.1016/j.canlet.2018.05.017
A Halama, H Kahal, A M Bhagwat, J Zierer, T Sathyapalan, J Graumann, K Suhre, S L Atkin (2018). Metabolic and proteomics signatures of hypoglycaemia in type 2 diabetes. Diabetes, obesity and metabolism, https://doi.org/10.1111/dom.13602