1 Introduction
2 Installation
3 Assumed output directory structure
4 Check that all required alevin files are available
5 Generate QC report
6 Create shiny app
7 Generate individual plots
8 Session info
References

1 Introduction

The purpose of the alevinQC package is to generate a summary QC report based on the output of an alevin (Srivastava et al. 2018) run. The QC report can be generated as a html or pdf file, or launched as a shiny application.

2 Installation

alevinQC can be installed using the BiocManager CRAN package.

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("alevinQC")

After installation, load the package into the R session.

library(alevinQC)

Note that in order to process output from Salmon v0.14 or later, you need Alevin v1.1 or later.

3 Assumed output directory structure

For more information about running alevin, we refer to the documentation. When invoked, alevin generates several output files in the specified output directory. alevinQC assumes that this structure is retained, and will return an error if it isn’t - thus, it is not recommended to move or rename the output files from alevin. alevinQC assumes that the following files (in the indicated structure) are available in the provided baseDir (note that currently, in order to generate the full set of files, alevin must be invoked with the --dumpFeatures flag).

baseDir
  |--alevin
  |    |--featureDump.txt
  |    |--filtered_cb_frequency.txt
  |    |--MappedUmi.txt
  |    |--quants_mat_cols.txt
  |    |--quants_mat_rows.txt
  |    |--quants_mat.gz
  |    |--raw_cb_frequency.txt
  |    |--whitelist.txt
  |--aux_info
  |    |--meta_info.json
  |--cmd_info.json

4 Check that all required alevin files are available

The report generation functions (see below) will check that all the required files are available in the provided base directory. However, you can also call the function checkAlevinInputFiles() to run the check manually. If one or more files are missing, the function will raise an error indicating the missing file(s).

baseDir <- system.file("extdata/alevin_example_v0.14", package = "alevinQC")
checkAlevinInputFiles(baseDir = baseDir)
#> [1] "v0.14"

5 Generate QC report

The alevinQCReport() function generates the QC report from the alevin output. Depending on the file extension of the outputFile argument, and the value of outputFormat, the function can generate either an html report or a pdf report.

outputDir <- tempdir()
alevinQCReport(baseDir = baseDir, sampleId = "testSample", 
               outputFile = "alevinReport.html", 
               outputFormat = "html_document",
               outputDir = outputDir, forceOverwrite = TRUE)

6 Create shiny app

In addition to static reports, alevinQC can also generate a shiny application, containing the same summary figures as the pdf and html reports.

app <- alevinQCShiny(baseDir = baseDir, sampleId = "testSample")

Once created, the app can be launched using the runApp() function from the shiny package.

shiny::runApp(app)

7 Generate individual plots

The individual plots included in the QC reports can also be independently generated. To do so, we must first read the alevin output into an R object.

alevin <- readAlevinQC(baseDir = baseDir)

The resulting list contains three entries:

cbTable: a data.frame with various inferred characteristics of the individual cell barcodes
summaryTables: a list of data.frames with summary information about the full data set, the initial set of whitelisted cells and the final set of whitelisted cells, respectively
versionTable: a matrix with information about the invokation of alevin

head(alevin$cbTable)
#>                 CB originalFreq ranking collapsedFreq nbrMappedUMI
#> 1 GACTGCGAGGGCATGT       121577       1        123419       104128
#> 2 GGTGCGTAGGCTACGA       110467       2        111987        93608
#> 3 ATGAGGGAGTAGTGCG       106446       3        108173        88481
#> 4 ACTGTCCTCATGCTCC       104794       4        106085        81879
#> 5 CGAACATTCTGATACG       104616       5        106072        84395
#> 6 ACTGTCCCATATGGTC        99208       6        100776        81066
#>   totalUMICount mappingRate dedupRate  MeanByMax nbrGenesAboveZero
#> 1         73312    0.843695  0.295943 0.00735194              7512
#> 2         66002    0.835883  0.294911 0.00783094              7522
#> 3         62196    0.817958  0.297069 0.00832595              7081
#> 4         57082    0.771824  0.302849 0.00619664              6956
#> 5         58547    0.795639  0.306274 0.00743685              7347
#> 6         56534    0.804418  0.302618 0.00947029              6841
#>   nbrGenesAboveMean ArborescenceCount inFinalWhiteList inFirstWhiteList
#> 1              1237           1.42034             TRUE             TRUE
#> 2              1238           1.41826             TRUE             TRUE
#> 3              1151           1.42262             TRUE             TRUE
#> 4               957           1.43441             TRUE             TRUE
#> 5              1238           1.44149             TRUE             TRUE
#> 6              1068           1.43393             TRUE             TRUE

knitr::kable(alevin$summaryTables$fullDataset)

Total number of processed reads	7197662
Number of reads with Ns	35362
Number of reads with valid cell barcode (no Ns)	7162300
Number of noisy CB reads	1003624
Number of noisy UMI reads	266
Total number of observed cell barcodes	188613

knitr::kable(alevin$summaryTables$initialWhitelist)

Number of barcodes in initial whitelist	100
Fraction reads in initial whitelist barcodes	84.64%
Mean number of reads per cell (initial whitelist)	60620
Median number of reads per cell (initial whitelist)	58132
Median number of detected genes per cell (initial whitelist)	5268
Median UMI count per cell (initial whitelist)	31353

knitr::kable(alevin$summaryTables$finalWhitelist)

Number of barcodes in final whitelist	95
Fraction reads in final whitelist barcodes	82.39%
Mean number of reads per cell (final whitelist)	62118
Median number of reads per cell (final whitelist)	58725
Median number of detected genes per cell (final whitelist)	5343
Median UMI count per cell (final whitelist)	32028

knitr::kable(alevin$versionTable)

Start time	Thu May 30 13:06:55 2019
Salmon version	0.14.0
Index	/mnt/scratch5/avi/alevin/data/mohu/salmon_index
R1file	/mnt/scratch5/avi/alevin/data/10x/v2/mohu/100/all_bcs.fq
R2file	/mnt/scratch5/avi/alevin/data/10x/v2/mohu/100/all_reads.fq
tgMap	/mnt/scratch5/avi/alevin/data/mohu/gtf/txp2gene.tsv

The plots can now be generated using the dedicated plotting functions provided with alevinQC (see the help file for the respective function for more information).

plotAlevinKneeRaw(alevin$cbTable)

plotAlevinBarcodeCollapse(alevin$cbTable)

plotAlevinQuant(alevin$cbTable)

plotAlevinKneeNbrGenes(alevin$cbTable)

8 Session info

sessionInfo()
#> R version 3.6.1 (2019-07-05)
#> Platform: x86_64-pc-linux-gnu (64-bit)
#> Running under: Ubuntu 18.04.3 LTS
#> 
#> Matrix products: default
#> BLAS:   /home/biocbuild/bbs-3.10-bioc/R/lib/libRblas.so
#> LAPACK: /home/biocbuild/bbs-3.10-bioc/R/lib/libRlapack.so
#> 
#> locale:
#>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
#>  [3] LC_TIME=en_US.UTF-8        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       
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] alevinQC_1.2.0   BiocStyle_2.14.0
#> 
#> loaded via a namespace (and not attached):
#>  [1] Rcpp_1.0.2           highr_0.8            later_1.0.0         
#>  [4] pillar_1.4.2         compiler_3.6.1       BiocManager_1.30.9  
#>  [7] RColorBrewer_1.1-2   plyr_1.8.4           tools_3.6.1         
#> [10] digest_0.6.22        evaluate_0.14        tibble_2.1.3        
#> [13] gtable_0.3.0         pkgconfig_2.0.3      rlang_0.4.1         
#> [16] shiny_1.4.0          GGally_1.4.0         crosstalk_1.0.0     
#> [19] yaml_2.2.0           xfun_0.10            fastmap_1.0.1       
#> [22] dplyr_0.8.3          stringr_1.4.0        knitr_1.25          
#> [25] htmlwidgets_1.5.1    shinydashboard_0.7.1 cowplot_1.0.0       
#> [28] DT_0.9               grid_3.6.1           tidyselect_0.2.5    
#> [31] reshape_0.8.8        glue_1.3.1           R6_2.4.0            
#> [34] rmarkdown_1.16       bookdown_0.14        purrr_0.3.3         
#> [37] ggplot2_3.2.1        magrittr_1.5         promises_1.1.0      
#> [40] scales_1.0.0         htmltools_0.4.0      tximport_1.14.0     
#> [43] assertthat_0.2.1     xtable_1.8-4         mime_0.7            
#> [46] colorspace_1.4-1     httpuv_1.5.2         labeling_0.3        
#> [49] stringi_1.4.3        lazyeval_0.2.2       munsell_0.5.0       
#> [52] rjson_0.2.20         crayon_1.3.4

References

Srivastava, Avi, Laraib Malik, Tom Sean Smith, Ian Sudbery, and Rob Patro. 2018. “Alevin Efficiently Estimates Accurate Gene Abundances from dscRNA-seq Data.” bioRxiv Doi:10.1101/335000.

alevinQC

2019-10-29

Abstract

Package

Contents