1 New samples!

In this copy of the worksheet, I am going to deal only with the tuberculist libraries.

This worksheet is a copy of 20180806 but this time with feeling! (or new samples) A minor caveat: I had some difficulties converting the new data to a openMS suitable format. I am not certain if I messed up a priori, or if a change in the version of the software on the spec caused the problem, but for future reference:

To properly convert the data to mzXML, make sure to use the TPP MSConvert utility and have ‘TPP’ compatibility on.

2 Analyzing data from openMS and friends.

2.1 SWATH2stats preprocessing

I am using my slightly modified copy of SWATH2stats. This seeks to ensure that changes in the case of columns in the metadata from one version of OpenMS to another do not trouble me.

I am separately performing the analysis for our home-grown libraries and the publicly available ones (documented in the 201806 worksheets).

2.1.1 Creating a swath2stats experiment using the tuberculist-derived library data

This just repeats the previous stanza, but uses the tuberculist transition-library swath data as its input rather than our own comet data.

There is one important caveat in the following block: I used a regex to remove the second half of geneID_geneName so that later when I merge in the annotation data I have it will match.

tb_tric_data <- read.csv(paste0("results/tric/", ver, "/whole_8mz_tuberculist/comet_HCD.tsv"),
                         sep="\t")
tb_tric_data[["ProteinName"]] <- gsub(pattern="^(.*)_.*$", replacement="\\1",
                                      x=tb_tric_data[["ProteinName"]])
sample_annot <- extract_metadata(paste0("sample_sheets/Mtb_dia_samples_", ver, ".xlsx"))
kept <- ! grepl(x=rownames(sample_annot), pattern="^s\\.\\.")
sample_annot <- sample_annot[kept, ]
devtools::load_all("~/scratch/git/SWATH2stats")
## Loading SWATH2stats
tb_s2s_exp <- sample_annotation(data=tb_tric_data, check_files=FALSE,
                                sample_annotation=sample_annot,
                                fullpeptidename_column="fullunimodpeptidename")
## Not checking that the files are identical between the annotation and data.
## Warning in sample_annotation(data = tb_tric_data, check_files = FALSE,
## sample_annotation = sample_annot, : No measurement value found for sample
## 'results/01mzXML/dia/20180806/2018_0726Briken03.mzXML' in the data file.
## Warning in sample_annotation(data = tb_tric_data, check_files = FALSE,
## sample_annotation = sample_annot, : No measurement value found for sample
## 'results/01mzXML/dia/20180806/2018_0726Briken11.mzXML' in the data file.
## Warning in sample_annotation(data = tb_tric_data, check_files = FALSE,
## sample_annotation = sample_annot, : No measurement value found for sample
## 'results/mzXML/dia/20180611/2018_0116BrikenDIA01.mzXML' in the data file.

Now I have a couple data structures which should prove useful for the metrics provided by SWATH2stats, MSstats, and my own hpgltools.

3 SWATH2stats continued

The various metrics and filters provided by SWATH2stats seem quite reasonable to me. The only thing that really bothers me is that they are all case sensitive and I found that the most recent tric changed the capitalization of a column, causing these to all fall down. Therefore I went in and made everything case insensitive in a fashion similar to that done by MSstats (except I hate capital letters, so I used tolower() rather than toupper()).

3.1 Run filters on the tuberculist

Repeat the above stanza using the tuberculist data.

## Get correlations on a sample by sample basis
pp(file="images/20180611_tb_swath2stats_sample_cor.png")
## Going to write the image to: images/20180611_tb_swath2stats_sample_cor.png when dev.off() is called.
sample_cor <- plot_correlation_between_samples(tb_s2s_exp, size=2,
                                               fun.aggregate=sum,
                                               column.values="intensity")
dev.off()
## png 
##   2
sample_cond_rep_cor <- plot_correlation_between_samples(tb_s2s_exp, size=2,
                                                        comparison=transition_group_id ~
                                                          condition + bioreplicate + run,
                                                        fun.aggregate=sum,
                                                        column.values="intensity")

decoy_lists <- assess_decoy_rate(tb_s2s_exp)
## Number of non-decoy peptides: 25277
## Number of decoy peptides: 2177
## Decoy rate: 0.0861
## This seems a bit high to me, yesno?
fdr_overall <- assess_fdr_overall(tb_s2s_exp, output="Rconsole", plot=TRUE)

byrun_fdr <- assess_fdr_byrun(tb_s2s_exp, FFT=0.7, plot=TRUE, output="Rconsole")
## The average FDR by run on assay level is 0.008
## The average FDR by run on peptide level is 0.009
## The average FDR by run on protein level is 0.039

chosen_mscore <- mscore4assayfdr(tb_s2s_exp, FFT=0.7, fdr_target=0.02)
## Target assay FDR: 0.02
## Required overall m-score cutoff: 0.0039811
## achieving assay FDR: 0.0189
prot_score <- mscore4protfdr(tb_s2s_exp, FFT=0.7, fdr_target=0.02)
## Target protein FDR: 0.02
## Required overall m-score cutoff: 0.00044668
## achieving protein FDR: 0.0188
tb_mscore_filtered <- filter_mscore(tb_s2s_exp, chosen_mscore)
## Original dimension: 532228, new dimension: 479136, difference: 53092.
tb_freq_mscore <- filter_mscore_freqobs(tb_s2s_exp, 0.01, 0.8, rm.decoy=FALSE)
## Peptides need to have been quantified in more conditions than: 45.6 in order to pass this percentage-based threshold.
## Fraction of peptides selected: 0.048
## Original dimension: 539476, new dimension: 82454, difference: 457022.
tb_data_filtered_fdr <- filter_mscore_fdr(tb_mscore_filtered, FFT=0.7,
                                          overall_protein_fdr_target=prot_score,
                                          upper_overall_peptide_fdr_limit=0.05)
## Target protein FDR: 0.000446683592150963
## Required overall m-score cutoff: 0.01
## achieving protein FDR: 0
## filter_mscore_fdr is filtering the data...
## finding m-score cutoff to achieve desired protein FDR in protein master list..
## finding m-score cutoff to achieve desired global peptide FDR..
## Target peptide FDR: 0.05
## Required overall m-score cutoff: 0.01
## Achieving peptide FDR: 0
## Proteins selected: 
## Total proteins selected: 3107
## Final target proteins: 3107
## Final decoy proteins: 0
## Peptides mapping to these protein entries selected:
## Total mapping peptides: 23563
## Final target peptides: 23563
## Final decoy peptides: 0
## Total peptides selected from:
## Total peptides: 23563
## Final target peptides: 23563
## Final decoy peptides: 0
## Individual run FDR quality of the peptides was not calculated
## as not every run contains a decoy.
## The decoys have been removed from the returned data.
tb_only_proteotypic <- filter_proteotypic_peptides(tb_data_filtered_fdr)
## Number of proteins detected: 3120
## Protein identifiers: Rv1270c, Rv0161, Rv1613, Rv2540c, Rv0440, Rv1327c
## Number of proteins detected that are supported by a proteotypic peptide: 2982
## Number of proteotypic peptides detected: 23398
tb_all_filtered <- filter_all_peptides(tb_only_proteotypic)
## Number of proteins detected: 2984
## First 6 protein identifiers: Rv1270c, Rv0161, Rv1613, Rv2540c, Rv0440, Rv1327c
tb_only_strong <- filter_on_max_peptides(data=tb_all_filtered, n_peptides=10)
## Before filtering:
##   Number of proteins: 2982
##   Number of peptides: 23398
## 
## Percentage of peptides removed: 23.68%
## 
## After filtering:
##   Number of proteins: 2960
##   Number of peptides: 17857
tb_only_minimum <- filter_on_min_peptides(data=tb_only_strong, n_peptides=3)
## Before filtering:
##   Number of proteins: 2960
##   Number of peptides: 17857
## 
## Percentage of peptides removed: 0%
## 
## After filtering:
##   Number of proteins: 2774
##   Number of peptides: 17857
## I think these matrixes are probably smarter to use than the raw outmatrix from tric.
## But I am not a fan of rerwriting the sample column names.
matrix_prefix <- file.path("results", "swath2stats", ver)
if (!file.exists(matrix_prefix)) {
  dir.create(matrix_prefix)
}
protein_matrix_all <- write_matrix_proteins(
  tb_s2s_exp, write.csv=TRUE,
  filename=file.path(matrix_prefix, "tb_protein_all.csv"))
## Protein overview matrix results/swath2stats/20180817/tb_protein_all.csv written to working folder.
dim(protein_matrix_all)
## [1] 4858   58
protein_matrix_mscore <- write_matrix_proteins(
  tb_mscore_filtered, write.csv=TRUE,
  filename=file.path(matrix_prefix, "tb_protein_matrix_mscore.csv"))
## Protein overview matrix results/swath2stats/20180817/tb_protein_matrix_mscore.csv written to working folder.
dim(protein_matrix_mscore)
## [1] 3107   58
peptide_matrix_mscore <- write_matrix_peptides(
  tb_mscore_filtered, write.csv=TRUE,
  filename=file.path(matrix_prefix, "tb_peptide_matrix_mscore.csv"))
## Peptide overview matrix results/swath2stats/20180817/tb_peptide_matrix_mscore.csv written to working folder.
dim(peptide_matrix_mscore)
## [1] 23563    58
protein_matrix_minimum <- write_matrix_proteins(
  tb_only_minimum, write.csv=TRUE,
  filename=file.path(matrix_prefix, "tb_protein_matrix_minimum.csv"))
## Protein overview matrix results/swath2stats/20180817/tb_protein_matrix_minimum.csv written to working folder.
dim(protein_matrix_minimum)
## [1] 2774   58
peptide_matrix_minimum <- write_matrix_peptides(
  tb_only_minimum, write.csv=TRUE,
  filename=file.path(matrix_prefix, "tb_peptide_matrix_minimum.csv"))
## Peptide overview matrix results/swath2stats/20180817/tb_peptide_matrix_minimum.csv written to working folder.
dim(peptide_matrix_minimum)
## [1] 345283     58
rt_cor <- plot_correlation_between_samples(
  tb_only_minimum, column.values="intensity", fun.aggregate=sum, size=2)

## I have no effing clue what this plot means.
variation <- plot_variation(tb_only_minimum, fun.aggregate=sum)

cols <- colnames(tb_all_filtered)
tb_disaggregated <- disaggregate(tb_all_filtered, all.columns=TRUE)
## The library contains5transitions per precursor.
## The data table was transformed into a table containing one row per transition.
tb_msstats_input <- convert_MSstats(tb_disaggregated)
## One or several columns required by MSstats were not in the data. The columns were created and filled with NAs.
## Missing columns: productcharge, isotopelabeltype
## isotopelabeltype was filled with light.

3.2 Some new plots

In response to some interesting queries from Yan, I made a few little functions which query and plot data from the scored data provided by openswath/pyprophet. Let us look at their results here.

3.2.1 The tuberculist data

tb_pyprophet_fun <- sm(extract_pyprophet_data(
  metadata=paste0("sample_sheets/Mtb_dia_samples_", ver, ".xlsx"),
  pyprophet_column="tuberculistscored"))

tb_mass_plot <- sm(plot_pyprophet_distribution(tb_pyprophet_fun, column="mass"))
tb_mass_plot[["violin"]]

tb_deltart_plot_all <- sm(plot_pyprophet_distribution(
  tb_pyprophet_fun, column="delta_rt"))
tb_deltart_plot_all[["violin"]]

tb_deltart_plot_real <- sm(plot_pyprophet_distribution(
  tb_pyprophet_fun,
  column="delta_rt", keep_decoys=FALSE))
tb_deltart_plot_real[["violin"]]

tb_deltart_plot_decoys <- sm(plot_pyprophet_distribution(
  tb_pyprophet_fun,
  column="delta_rt", keep_real=FALSE))
tb_deltart_plot_decoys[["violin"]]

3.3 MSstats

msstats.org seems to provide a complete solution for performing reasonable metrics of this data.

I am currently reading: http://msstats.org/wp-content/uploads/2017/01/MSstats_v3.7.3_manual.pdf

I made some moderately intrusive changes to MSstats to make it clearer, as well.

3.3.1 Tuberculist derived msstats

tt <- sm(devtools::load_all("~/scratch/git/MSstats"))
checkpoint <- "tb_dataprocess.rda"
if (file.exists(checkpoint)) {
  load(file=checkpoint)
} else {
  tb_msstats_quant <- sm(dataProcess(tb_msstats_input))
  save(file=checkpoint, list=c("tb_msstats_quant"))
}
checkpoint <- "tb_plots.rda"
if (file.exists(checkpoint)) {
  load(file=checkpoint)
} else {
  tb_msstats_plots <- sm(dataProcessPlots(tb_msstats_quant, type="QCPLOT"))
  save(file=checkpoint, list=c("tb_msstats_plots"))
}

my_levels <- levels(as.factor(tb_msstats_input$condition))
my_levels
## [1] "wt_whole"       "wt_filtrate"    "delta_whole"    "comp_whole"    
## [5] "delta_filtrate" "comp_filtrate"
comparisons <- ghetto_contrast_matrix(
  numerators=c("wt_filtrate", "delta_filtrate", "comp_filtrate",
               "delta_filtrate", "comp_filtrate", "delta_whole",
               "comp_whole"),
  denominators=c("wt_whole", "delta_whole", "comp_whole",
                 "wt_filtrate", "wt_filtrate", "wt_whole",
                 "wt_whole"))

tb_results <- list()
checkpoint <- "tb_group.rda"
if (file.exists(checkpoint)) {
  load(file=checkpoint)
} else {
  for (c in 1:length(rownames(comparisons))) {
    name <- rownames(comparisons)[c]
    message("Starting ", name)
    comp <- comparisons[c, ]
    comp <- t(as.matrix(comp))
    rownames(comp) <- name
    tb_results[[name]] <- sm(MSstats::groupComparison(contrast.matrix=comp,
                                                      data=tb_msstats_quant))
    message("Finished ", name)
  }
  save(file=checkpoint, list=c("tb_results"))
}

3.3.2 P/PE protein QC plots for Yan

Yan asked for the p/pe protein qc plots. ok. I changed the dataProcessPlots to return something useful, so that should be possible now.

pe_genes <- read.table("reference/annotated_pe_genes.txt")[[1]]

## Unfortunately, the names did not get set in my changed version of dataProcessPlots...
plotlst <- our_msstats_plots$QCPLOT
## Error in eval(expr, envir, enclos): object 'our_msstats_plots' not found
available_plots <- gsub(pattern="^1/", replacement="",
                        x=levels(our_msstats_quant$ProcessedData$PROTEIN))
## Error in levels(our_msstats_quant$ProcessedData$PROTEIN): object 'our_msstats_quant' not found
names(plotlst) <- available_plots
## Error in eval(expr, envir, enclos): object 'available_plots' not found
pe_in_avail_idx <- pe_genes %in% available_plots
## Error in pe_genes %in% available_plots: object 'available_plots' not found
pe_in_avail <- pe_genes[pe_in_avail_idx]
## Error in eval(expr, envir, enclos): object 'pe_in_avail_idx' not found
pe_plots <- plotlst[pe_in_avail]
## Error in eval(expr, envir, enclos): object 'plotlst' not found
pdf(file="pe_qc_plots.pdf")
for (p in 1:length(pe_plots)) {
  plot(pe_plots[[p]])
}
## Error in eval(expr, envir, enclos): object 'pe_plots' not found
dev.off()
## png 
##   2
length(pe_plots)
## Error in eval(expr, envir, enclos): object 'pe_plots' not found

4 Create hpgltools expressionset

Since I am not certain I understand these data, I will take the intensities from SWATH2stats, metadata, and annotation data; attempt to create a ‘normal’ expressionset; poke at it to see what I can learn.

4.1 Massaging the metadata

I want to use the same metadata as were used for MSstats. It has a few important differences from the requirements of hpgltools: pretty much only that I do not allow rownames/sampleIDs to start with a number.

4.2 Massaging the intensity matrix

I do not want the \1 before the protein names, I already merged them into one entry per gene vis SWATH2stats.

4.2.1 Make matrices with the tb libraries

tb_prot_mtrx <- read.csv(file.path("results", "swath2stats", ver, "tb_protein_matrix_minimum.csv"))
rownames(tb_prot_mtrx) <- gsub(pattern="^1\\/", replacement="", x=tb_prot_mtrx[["proteinname"]])
tb_prot_mtrx <- tb_prot_mtrx[, -1]
## Important question: Did SWATH2stats reorder my data?
colnames(tb_prot_mtrx) <- gsub(pattern="^(.*)(2018.*)$", replacement="s\\2", x=colnames(tb_prot_mtrx))

4.3 Merge the pieces

Now we should have sufficient pieces to make an expressionset.

While here, I will also split the data into a cf and whole-cell pair of data structures.

4.3.1 Tb data

## Drop the metadata not in the protein matrix:
## And ensure that they are the same order.
reordered <- colnames(tb_prot_mtrx)
metadata <- sample_annot[reordered, ]

tb_protein_expt <- sm(create_expt(metadata,
                                  count_dataframe=tb_prot_mtrx,
                                  gene_info=mtb_annotations))
tb_whole_expt <- subset_expt(tb_protein_expt, subset="collectiontype=='whole'")
## There were 57, now there are 28 samples.
tb_cf_expt <- subset_expt(tb_protein_expt, subset="collectiontype=='filtrate'")
## There were 57, now there are 29 samples.

4.4 Metrics of the full data set

4.4.1 Tb libraries

tb_protein_metrics <- sm(graph_metrics(tb_protein_expt))
tb_protein_norm <- sm(normalize_expt(tb_protein_expt, transform="log2", convert="cpm",
                                      norm="quant", filter=TRUE))
tb_protein_norm_metrics <- sm(graph_metrics(tb_protein_norm))
tb_protein_fsva <- sm(normalize_expt(tb_protein_expt, transform="log2", convert="cpm",
                                      batch="fsva", filter=TRUE))
tb_protein_fsva_metrics <- sm(graph_metrics(tb_protein_fsva))

4.5 Metrics of the whole-cell data set

4.5.1 Tb libraries

tb_whole_metrics <- sm(graph_metrics(tb_whole_expt))
tb_whole_norm <- sm(normalize_expt(tb_whole_expt, transform="log2", convert="cpm",
                                    norm="quant", filter=TRUE))
tb_whole_norm_metrics <- sm(graph_metrics(tb_whole_norm))
tb_whole_fsva <- sm(normalize_expt(tb_whole_expt, transform="log2", convert="cpm",
                                    batch="fsva", filter=TRUE))
tb_whole_fsva_metrics <- sm(graph_metrics(tb_whole_fsva))
tb_whole_varpart <- varpart(tb_whole_expt)
## Attempting mixed linear model with: ~  (1|condition) + (1|batch)
## Fitting the expressionset to the model, this is slow.
## Projected run time: ~ 0.2 min
## Placing factor: condition at the beginning of the model.
tb_whole_varpart$partition_plot

tb_whole_cv <- plot_variance_coefficients(tb_protein_expt)
## Naively calculating coefficient of variation and quartile dispersion with respect to condition.
## Finished calculating dispersion estimates.
tb_whole_cv$disp

tb_batch_cv <- plot_variance_coefficients(tb_protein_expt, x_axis="batch")
## Naively calculating coefficient of variation and quartile dispersion with respect to batch.
## Finished calculating dispersion estimates.
tb_batch_cv$disp

tb_genotype_cv <- plot_variance_coefficients(tb_protein_expt, x_axis="genotype")
## Naively calculating coefficient of variation and quartile dispersion with respect to genotype.
## Finished calculating dispersion estimates.
tb_genotype_cv$disp

tb_prep_cv <- plot_variance_coefficients(tb_protein_expt, x_axis="prepdate")
## Naively calculating coefficient of variation and quartile dispersion with respect to prepdate.
## Finished calculating dispersion estimates.
tb_prep_cv$disp

vio <- plot_boxplot(tb_protein_expt, violin=TRUE)
## This data will benefit from being displayed on the log scale.
## If this is not desired, set scale='raw'
## Some entries are 0.  We are on log scale, adding 1 to the data.
## Changed 74400 zero count features.

4.6 Metrics of the filtrate data set

4.6.1 Tb libraries

tb_cf_metrics <- sm(graph_metrics(tb_cf_expt))
tb_cf_norm <- sm(normalize_expt(tb_cf_expt, transform="log2", convert="cpm",
                                 norm="quant", filter=TRUE))
tb_cf_norm_metrics <- sm(graph_metrics(tb_cf_norm))
tb_cf_fsva <- sm(normalize_expt(tb_cf_expt, transform="log2", convert="cpm",
                                 batch="fsva", filter=TRUE))
tb_cf_fsva_metrics <- sm(graph_metrics(tb_cf_fsva))

4.7 plot some metrics

4.7.1 Tb libraries

pp(image=tb_protein_metrics$libsize,
   file=file.path("images", paste0(ver, "_tb_libsize.png")))
## Writing the image to: images/20180817_tb_libsize.png and calling dev.off().

## It seems to me that the scale of the data is all within an order of magnitude or two.
## I cannot get used to these absurdly large numbers though.
pp(image=tb_protein_norm_metrics$pcaplot,
   file=file.path("images", paste0(ver, "_tb_norm_pca.png")))
## Writing the image to: images/20180817_tb_norm_pca.png and calling dev.off().
## Warning in MASS::cov.trob(data[, vars]): Probable convergence failure

## Warning in MASS::cov.trob(data[, vars]): Probable convergence failure

## Warning in MASS::cov.trob(data[, vars]): Probable convergence failure

## Warning in MASS::cov.trob(data[, vars]): Probable convergence failure

## There appears to be a nice split in the data, however the un-assayable batch
## effect is a problem.
pp(image=tb_protein_fsva_metrics$pcaplot,
   file=file.path("images", paste0(ver, "_fsva_pca.png")))
## Writing the image to: images/20180817_fsva_pca.png and calling dev.off().

## fsva seems to get some handle on the data, but I don't think we should rely
## upon it.
pp(image=tb_protein_norm_metrics$corheat,
   file=file.path("images", paste0(ver, "_norm_corheat.png")))
## Writing the image to: images/20180817_norm_corheat.png and calling dev.off().

## Once again, the whole-cell/culture-filtrate split is very large.
pp(image=tb_protein_metrics$density,
   file=file.path("images", paste0(ver, "_raw_density.png")))
## Writing the image to: images/20180817_raw_density.png and calling dev.off().

## There are two obvious distributions in the data, once again split between types.
pp(image=tb_protein_metrics$boxplot,
   file=file.path("images", paste0(ver, "_boxplot.png")))
## Writing the image to: images/20180817_boxplot.png and calling dev.off().

## This recapitulates the previous plot.

pp(image=tb_whole_metrics$libsize,
   file=file.path("images", paste0(ver, "_whole_libsize.png")))
## Writing the image to: images/20180817_whole_libsize.png and calling dev.off().

pp(image=tb_whole_norm_metrics$pcaplot,
   file=file.path("images", paste0(ver, "_whole_norm_pca.png")))
## Writing the image to: images/20180817_whole_norm_pca.png and calling dev.off().

pp(image=tb_whole_fsva_metrics$pcaplot,
   file=file.path("images", paste0(ver, "_whole_fsva_pca.png")))
## Writing the image to: images/20180817_whole_fsva_pca.png and calling dev.off().

pp(image=tb_whole_norm_metrics$corheat,
   file=file.path("images", paste0(ver, "_whole_norm_corheat.png")))
## Writing the image to: images/20180817_whole_norm_corheat.png and calling dev.off().

pp(image=tb_whole_metrics$density,
   file=file.path("images", paste0(ver, "_whole_raw_density.png")))
## Writing the image to: images/20180817_whole_raw_density.png and calling dev.off().

pp(image=tb_whole_metrics$boxplot,
   file=file.path("images", paste0(ver, "_whole_boxplot.png")))
## Writing the image to: images/20180817_whole_boxplot.png and calling dev.off().

pp(image=tb_cf_metrics$libsize,
   file=file.path("images", paste0(ver, "_libsize.png")))
## Writing the image to: images/20180817_libsize.png and calling dev.off().

pp(image=tb_cf_norm_metrics$pcaplot,
   file=file.path("images", paste0(ver, "_norm_pca.png")))
## Writing the image to: images/20180817_norm_pca.png and calling dev.off().

pp(image=tb_cf_fsva_metrics$pcaplot,
   file=file.path("images", paste0(ver, "_fsva_pca.png")))
## Writing the image to: images/20180817_fsva_pca.png and calling dev.off().

pp(image=tb_cf_norm_metrics$corheat,
   file=file.path("images", paste0(ver, "_norm_corheat.png")))
## Writing the image to: images/20180817_norm_corheat.png and calling dev.off().

pp(image=tb_cf_metrics$density,
   file=file.path("images", paste0(ver, "_raw_density.png")))
## Writing the image to: images/20180817_raw_density.png and calling dev.off().

pp(image=tb_cf_metrics$boxplot,
   file=file.path("images", paste0(ver, "_boxplot.png")))
## Writing the image to: images/20180817_boxplot.png and calling dev.off().

5 Collapse technical replicates

I am thinking that a simple sum of the data in the technical replicates should be sufficient.

tb_compressed <- concatenate_runs(tb_protein_expt, column="bioreplicate")
## The original expressionset has 57 samples.
## The final expressionset has 18 samples.
tb_whole_compressed <- concatenate_runs(tb_whole_expt, column="bioreplicate")
## The original expressionset has 28 samples.
## The final expressionset has 9 samples.
tb_cf_compressed <- concatenate_runs(tb_cf_expt, column="bioreplicate")
## The original expressionset has 29 samples.
## The final expressionset has 9 samples.

5.1 Reperform metrics with the technical replicates removed.

tb_compressed_metrics <- sm(graph_metrics(tb_compressed))
tb_compressed_norm <- sm(normalize_expt(tb_compressed, transform="log2", convert="cpm",
                                         norm="quant", filter=TRUE))
tb_compressed_norm_metrics <- sm(graph_metrics(tb_compressed_norm))
tb_compressed_norm_metrics$pcaplot
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse
## Too few points to calculate an ellipse

tb_compressed_de <- sm(all_pairwise(tb_compressed, model_batch=FALSE, force=TRUE,
                                    do_ebseq=TRUE, parallel=FALSE))

5.1.1 Test my contrasts

I made some changes to my code recently which might have messed up the numerator/denominators in some (hopefully rare) cases. The following is a test to make sure that is caught now (as I noticed the error in this data set.

test_limma <- limma_pairwise(tb_compressed, model_batch=FALSE, force=TRUE)
test_deseq <- deseq_pairwise(tb_compressed, model_batch=FALSE, force=TRUE)
test_edger <- edger_pairwise(tb_compressed, model_batch=FALSE, force=TRUE)
test_basic <- basic_pairwise(tb_compressed, model_batch=FALSE, force=TRUE)
test_ebseq <- ebseq_pairwise(tb_compressed, model_batch=FALSE, force=TRUE)

tl <- test_limma$all_tables[[1]]
tb <- test_basic$all_tables[[1]]
tm <- merge(tl, tb, by="row.names")
cor.test(tm$logFC.x, tm$logFC.y)

tl <- test_limma$all_tables[[1]]
tb <- test_deseq$all_tables[[1]]
tm <- merge(tl, tb, by="row.names")
cor.test(tm$logFC.x, tm$logFC.y)

tl <- test_limma$all_tables[[1]]
tb <- test_edger$all_tables[[1]]
tm <- merge(tl, tb, by="row.names")
cor.test(tm$logFC.x, tm$logFC.y)

tl <- test_deseq$all_tables[[1]]
tb <- test_edger$all_tables[[1]]
tm <- merge(tl, tb, by="row.names")
cor.test(tm$logFC.x, tm$logFC.y)

tl <- test_limma$all_tables[[1]]
tb <- test_ebseq$all_tables[[1]]
tm <- merge(tl, tb, by="row.names")
cor.test(tm$logFC.x, tm$logFC.y)

6 Attempt some quantification comparisons?

6.1 Tb libraries

tb_pairwise_filt <- sm(normalize_expt(tb_compressed, filter=TRUE))
tb_pairwise_comp <- sm(all_pairwise(tb_pairwise_filt, model_batch="fsva", parallel=FALSE,
                                    do_ebseq=TRUE, force=TRUE))
tb_pairwise_nobatch <- sm(all_pairwise(tb_pairwise_filt, model_batch=FALSE, parallel=FALSE,
                                       do_ebseq=TRUE, force=TRUE))

6.1.1 Show a few metrics from the hpgltools pairwise comparisons

tb_compressed_de$comparison$heat
## NULL

7 For each msstats run, do a DE table

7.1 wt_cf vs wt_whole

7.1.1 Tb libraries

keepers <- list(
  "wtcf_vs_wtwhole" = c("wt_filtrate", "wt_whole"))
set_name <- "wt_filtrate_vs_wt_whole"
msstats_result <- tb_results[[set_name]][["ComparisonResult"]]
droppers <- c("undefined")
names(droppers) <- "log2fc"
## Make sure to set the rownames so it will merge into the excel file.
rownames(msstats_result) <- msstats_result[["Protein"]]
tb_wtcf_nobatch_wtwhole_tables <- sm(combine_de_tables(
  tb_pairwise_nobatch, keepers=keepers, extra_annot=msstats_result,
  excludes=droppers,
  excel=paste0("excel/tb_wtcf_vs_wtwhole_nobatch_tables-v", ver, ".xlsx")))
## Error in isTRUE(inverse): object 'inverse' not found
comp_table <- tb_wtcf_nobatch_wtwhole_tables$data[[set_name]]
## Error in eval(expr, envir, enclos): object 'tb_wtcf_nobatch_wtwhole_tables' not found
cor.test(comp_table$log2fc, comp_table$deseq_logfc, method="spearman")
## Error in cor.test(comp_table$log2fc, comp_table$deseq_logfc, method = "spearman"): object 'comp_table' not found

7.2 delta_cf vs delta_whole

7.2.1 Tb libraries

keepers <- list(
  "deltacf_vs_deltawhole" = c("delta_filtrate", "delta_whole"))
set_name <- "delta_filtrate_vs_delta_whole"
msstats_result <- tb_results[[set_name]][["ComparisonResult"]]
## Make sure to set the rownames so it will merge into the excel file.
rownames(msstats_result) <- msstats_result[["Protein"]]
tb_deltacf_deltawhole_tables <- sm(combine_de_tables(
  tb_pairwise_nobatch, keepers=keepers, extra_annot=msstats_result,
  excel=paste0("excel/tb_deltacf_vs_deltawhole_tables-v", ver, ".xlsx")))
## Error in isTRUE(inverse): object 'inverse' not found
comp_table <- tb_deltacf_deltawhole_tables$data[[set_name]]
## Error in eval(expr, envir, enclos): object 'tb_deltacf_deltawhole_tables' not found
cor.test(comp_table$log2fc, comp_table$deseq_logfc, method="spearman")
## Error in cor.test(comp_table$log2fc, comp_table$deseq_logfc, method = "spearman"): object 'comp_table' not found

7.3 comp_cf vs comp_whole

7.3.1 Tb libraries

keepers <- list(
  "compcf_vs_compwhole" = c("comp_filtrate", "comp_whole"))
set_name <- "comp_filtrate_vs_comp_whole"
msstats_result <- tb_results[[set_name]][["ComparisonResult"]]
## Make sure to set the rownames so it will merge into the excel file.
rownames(msstats_result) <- msstats_result[["Protein"]]
compcf_compwhole_tables <- sm(combine_de_tables(
  tb_pairwise_nobatch, keepers=keepers, extra_annot=msstats_result,
  excel=paste0("excel/tb_compcf_vs_compwhole_tables-v", ver, ".xlsx")))
## Error in isTRUE(inverse): object 'inverse' not found
comp_table <- compcf_compwhole_tables$data[[set_name]]
## Error in eval(expr, envir, enclos): object 'compcf_compwhole_tables' not found
cor.test(comp_table$log2fc, comp_table$deseq_logfc, method="spearman")
## Error in cor.test(comp_table$log2fc, comp_table$deseq_logfc, method = "spearman"): object 'comp_table' not found

7.4 delta_cf vs wt_cf

7.4.1 Tb libraries

keepers <- list(
  "deltacf_vs_wtcf" = c("delta_filtrate", "wt_filtrate"))
set_name <- "delta_filtrate_vs_wt_filtrate"
msstats_result <- tb_results[[set_name]][["ComparisonResult"]]
rownames(msstats_result) <- msstats_result[["Protein"]]
deltacf_wtcf_tables <- sm(combine_de_tables(
  tb_pairwise_nobatch, keepers=keepers, extra_annot=msstats_result,
  excel=paste0("excel/tb_deltacf_vs_wtcf_tables-v", ver, ".xlsx")))
## Error in isTRUE(inverse): object 'inverse' not found
comp_table <- deltacf_wtcf_tables$data[[set_name]]
## Error in eval(expr, envir, enclos): object 'deltacf_wtcf_tables' not found
cor.test(comp_table$log2fc, comp_table$deseq_logfc, method="spearman")
## Error in cor.test(comp_table$log2fc, comp_table$deseq_logfc, method = "spearman"): object 'comp_table' not found

7.5 comp_cf vs wt_cf

7.5.1 Tb libraries

keepers <- list(
  "compcf_vs_wtcf" = c("comp_filtrate", "wt_filtrate"))
set_name <- "comp_filtrate_vs_wt_filtrate"
msstats_result <- tb_results[[set_name]][["ComparisonResult"]]
rownames(msstats_result) <- msstats_result[["Protein"]]
compcf_wtcf_tables <- sm(combine_de_tables(
  tb_pairwise_nobatch, keepers=keepers, extra_annot=msstats_result,
  excel=paste0("excel/tb_compcf_vs_wtcf_tables-v", ver, ".xlsx")))
## Error in isTRUE(inverse): object 'inverse' not found
comp_table <- compcf_wtcf_tables$data[[set_name]]
## Error in eval(expr, envir, enclos): object 'compcf_wtcf_tables' not found
cor.test(comp_table$log2fc, comp_table$deseq_logfc, method="spearman")
## Error in cor.test(comp_table$log2fc, comp_table$deseq_logfc, method = "spearman"): object 'comp_table' not found

7.6 delta_whole vs wt_whole

7.6.1 Tb libraries

keepers <- list(
  "wtcf_vs_wtwhole" = c("wt_filtrate", "wt_whole"))
set_name <- "wt_filtrate_vs_wt_whole"
msstats_result <- tb_results[[set_name]][["ComparisonResult"]]
rownames(msstats_result) <- msstats_result[["Protein"]]
wtcf_wtwhole_tables <- sm(combine_de_tables(
  tb_pairwise_nobatch, keepers=keepers, extra_annot=msstats_result,
  excel=paste0("excel/tb_deltawhole_vs_wtwhole_tables-v", ver, ".xlsx")))
## Error in isTRUE(inverse): object 'inverse' not found
comp_table <- wtcf_wtwhole_tables$data[[set_name]]
## Error in eval(expr, envir, enclos): object 'wtcf_wtwhole_tables' not found
cor.test(comp_table$log2fc, comp_table$deseq_logfc, method="spearman")
## Error in cor.test(comp_table$log2fc, comp_table$deseq_logfc, method = "spearman"): object 'comp_table' not found

7.7 comp_whole vs wt_whole

7.7.1 Tb libraries

keepers <- list(
  "compwhole_vs_wtwhole" = c("comp_whole", "wt_whole"))
set_name <- "comp_whole_vs_wt_whole"
msstats_result <- tb_results[[set_name]][["ComparisonResult"]]
rownames(msstats_result) <- msstats_result[["Protein"]]
compwhole_wtwhole_tables <- sm(combine_de_tables(
  tb_pairwise_comp, keepers=keepers, extra_annot=msstats_result,
  excel=paste0("excel/tb_compwhole_vs_wtwhole_tables-v", ver, ".xlsx")))
## Error in isTRUE(inverse): object 'inverse' not found
comp_table <- compwhole_wtwhole_tables$data[[set_name]]
## Error in eval(expr, envir, enclos): object 'compwhole_wtwhole_tables' not found
cor.test(comp_table$log2fc, comp_table$deseq_logfc, method="spearman")
## Error in cor.test(comp_table$log2fc, comp_table$deseq_logfc, method = "spearman"): object 'comp_table' not found

7.8 Venn of MSstats vs. others for wt whole/cf

Najib asked for the overlap in perceived significance.

start <- wtcf_wtwhole_tables[["data"]][[1]]
## Error in eval(expr, envir, enclos): object 'wtcf_wtwhole_tables' not found
na_idx <- is.na(start[["adjpvalue"]])
## Error in start[["adjpvalue"]]: object of type 'closure' is not subsettable
start[na_idx, "adjpvalue"] <- 1
## Error in `[<-`(`*tmp*`, na_idx, "adjpvalue", value = 1): object 'na_idx' not found
ms_sig_idx <- start[["adjpvalue"]] <= 0.05
## Error in start[["adjpvalue"]]: object of type 'closure' is not subsettable
de_sig_idx <- start[["deseq_adjp"]] <= 0.05
## Error in start[["deseq_adjp"]]: object of type 'closure' is not subsettable
ed_sig_idx <- start[["edger_adjp"]] <= 0.05
## Error in start[["edger_adjp"]]: object of type 'closure' is not subsettable
lm_sig_idx <- start[["limma_adjp"]] <= 0.05
## Error in start[["limma_adjp"]]: object of type 'closure' is not subsettable
eb_sig_idx <- start[["ebseq_ppee"]] <= 0.05
## Error in start[["ebseq_ppee"]]: object of type 'closure' is not subsettable
ms_sig <- start[ms_sig_idx, ]
## Error in start[ms_sig_idx, ]: object 'ms_sig_idx' not found
de_sig <- start[de_sig_idx, ]
## Error in start[de_sig_idx, ]: object 'de_sig_idx' not found
ed_sig <- start[ed_sig_idx, ]
## Error in start[ed_sig_idx, ]: object 'ed_sig_idx' not found
lm_sig <- start[lm_sig_idx, ]
## Error in start[lm_sig_idx, ]: object 'lm_sig_idx' not found
eb_sig <- start[eb_sig_idx, ]
## Error in start[eb_sig_idx, ]: object 'eb_sig_idx' not found
ms_de_shared <- sum(rownames(ms_sig) %in% rownames(de_sig))
## Error in rownames(ms_sig): object 'ms_sig' not found
ms_solo <- ! rownames(ms_sig) %in% rownames(de_sig)
## Error in rownames(ms_sig): object 'ms_sig' not found
de_solo <- ! rownames(de_sig) %in% rownames(ms_sig)
## Error in rownames(de_sig): object 'de_sig' not found
ms_de_weights <- c(0, sum(ms_solo), sum(de_solo), sum(ms_de_shared))
## Error in eval(expr, envir, enclos): object 'ms_solo' not found
ms_de_venn <- Vennerable::Venn(SetNames=c("ms", "de"), Weight=ms_de_weights)
## Error in Vennerable::Venn(SetNames = c("ms", "de"), Weight = ms_de_weights): object 'ms_de_weights' not found
pp(file="/tmp/ms_de_venn.png")
## Going to write the image to: /tmp/ms_de_venn.png when dev.off() is called.
plot(ms_de_venn, doWeights=FALSE)
## Error in plot(ms_de_venn, doWeights = FALSE): object 'ms_de_venn' not found
dev.off()
## png 
##   2
ms_ed_shared <- sum(rownames(ms_sig) %in% rownames(ed_sig))
## Error in rownames(ms_sig): object 'ms_sig' not found
ms_solo <- ! rownames(ms_sig) %in% rownames(ed_sig)
## Error in rownames(ms_sig): object 'ms_sig' not found
ed_solo <- ! rownames(ed_sig) %in% rownames(ms_sig)
## Error in rownames(ed_sig): object 'ed_sig' not found
ms_ed_weights <- c(0, sum(ms_solo), sum(ed_solo), sum(ms_ed_shared))
## Error in eval(expr, envir, enclos): object 'ms_solo' not found
ms_ed_venn <- Vennerable::Venn(SetNames=c("ms", "ed"), Weight=ms_ed_weights)
## Error in Vennerable::Venn(SetNames = c("ms", "ed"), Weight = ms_ed_weights): object 'ms_ed_weights' not found
plot(ms_ed_venn, doWeights=FALSE)
## Error in plot(ms_ed_venn, doWeights = FALSE): object 'ms_ed_venn' not found
ms_lm_shared <- sum(rownames(ms_sig) %in% rownames(lm_sig))
## Error in rownames(ms_sig): object 'ms_sig' not found
ms_solo <- ! rownames(ms_sig) %in% rownames(lm_sig)
## Error in rownames(ms_sig): object 'ms_sig' not found
lm_solo <- ! rownames(lm_sig) %in% rownames(ms_sig)
## Error in rownames(lm_sig): object 'lm_sig' not found
ms_lm_weights <- c(0, sum(ms_solo), sum(lm_solo), sum(ms_lm_shared))
## Error in eval(expr, envir, enclos): object 'ms_solo' not found
ms_lm_venn <- Vennerable::Venn(SetNames=c("ms", "lm"), Weight=ms_lm_weights)
## Error in Vennerable::Venn(SetNames = c("ms", "lm"), Weight = ms_lm_weights): object 'ms_lm_weights' not found
plot(ms_lm_venn, doWeights=FALSE)
## Error in plot(ms_lm_venn, doWeights = FALSE): object 'ms_lm_venn' not found
ms_eb_shared <- sum(rownames(ms_sig) %in% rownames(eb_sig))
## Error in rownames(ms_sig): object 'ms_sig' not found
ms_solo <- ! rownames(ms_sig) %in% rownames(eb_sig)
## Error in rownames(ms_sig): object 'ms_sig' not found
eb_solo <- ! rownames(eb_sig) %in% rownames(ms_sig)
## Error in rownames(eb_sig): object 'eb_sig' not found
ms_eb_weights <- c(0, sum(ms_solo), sum(eb_solo), sum(ms_eb_shared))
## Error in eval(expr, envir, enclos): object 'ms_solo' not found
ms_eb_venn <- Vennerable::Venn(SetNames=c("ms", "eb"), Weight=ms_eb_weights)
## Error in Vennerable::Venn(SetNames = c("ms", "eb"), Weight = ms_eb_weights): object 'ms_eb_weights' not found
plot(ms_eb_venn, doWeights=FALSE)
## Error in plot(ms_eb_venn, doWeights = FALSE): object 'ms_eb_venn' not found

8 TODO

  • 2018-04-10: Make sure my invocations of SWATH2stats/MSstats are correct.
if (!isTRUE(get0("skip_load"))) {
  message(paste0("This is hpgltools commit: ", get_git_commit()))
  this_save <- paste0(gsub(pattern="\\.Rmd", replace="", x=rmd_file), "-v", ver, ".rda.xz")
  message(paste0("Saving to ", this_save))
  tmp <- sm(saveme(filename=this_save))
  pander::pander(sessionInfo())
}
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