Prepare analysis workflow

Set parameters

knitr::opts_knit$set(root.dir = rprojroot::find_rstudio_root_file(),
                     fig.width=15,
                     digit=5,
                     scipen=8)
options(digits=5, 
        scipen=8,
        future.globals.maxSize = +Inf)

Set filepaths

dataset_name <- commandArgs(trailingOnly=T)[1]
#dataset_name <- "hiseq4000"
message(sprintf("Dataset name: %s", dataset_name))
Dataset name: novaseq_l1
project_dir <- rprojroot::find_rstudio_root_file()

if(is.null(project_dir)){
  project_dir <- getwd()
  warning(sprintf("No rstudio project root file  found. 
                  Setting project directory to current workflow.Rmd file location: %s. 
                  Override if needed.",
                  project_dir))
 
}
message(sprintf("Project directory: %s",
                project_dir))
Project directory: /project/6007998/rfarouni/index_hopping

Load libraries

#library(DropletUtils)
library(tidyverse)
library(matrixStats)
library(broom)
library(furrr)
library(tictoc)
library(data.table)
library(cowplot)
library(rhdf5)
plan(multiprocess)

Load functions

code_dir <- file.path(project_dir, "code")
source(file.path(code_dir, "analysis_functions.R"))
source(file.path(code_dir, "io_functions.R"))
source(file.path(code_dir, "workflow_functions.R"))
source(file.path(code_dir, "plotting_functions.R"))

Run workflow

Estimate the sample index hopping probability, infer the true sample of origin, and find the optimal posterior probability threshold for retaining predicted real molecules.

max_fpr <- NULL # manually set the maximum false positive rate (not recommended)
data_list <- run_workflow(dataset_name, 
                          project_dir, 
                          max_fpr=max_fpr)
Step 1: reading read counts from existing file: 540.666 sec elapsed
Step 2: creating outcome counts datatable with grouping vars: 71.602 sec elapsed
Step 3: creating a chimera counts datatable and estimating hopping rate: 0.147 sec elapsed
Step 4: computing read counts distribution statistics: 0.237 sec elapsed
Step 5: estimating pi_r matrix: 0.032 sec elapsed
Step 6: infering the true sample of origin: 16.302 sec elapsed
Step 7: estimating g and computing classification metrics: 0.102 sec elapsed
Step 8: determining the optimal cutoff: 0.062 sec elapsed
Step 9: computing proportion of nonmissingness and updating summary data list: 0.011 sec elapsed
Step 10.1: reassigning reads to sample of origin: 3.156 sec elapsed
Step 10.2: deduplicating read counts: 0.021 sec elapsed
Step 10.3: reassigning hopped reads and deduplicating read counts: 3.201 sec elapsed
Step 10.4: labelling phantom molecules below cutoff: 0.001 sec elapsed
Step 10.5: adding cell-umi-gene labels: 77.54 sec elapsed
Step 10.6: tallying molecule counts by cell-barcode and gene ID: 48.49 sec elapsed
Step 10.7: tranforming cell-gene molecule tally table into long format: 418.564 sec elapsed
Step 10: purging phantoms at q cutoff of  0.960838. Max-FPR threshold user-set FALSE: 890.973 sec elapsed
Step 11: calling cells: 2105.52 sec elapsed
Step 12: saving purged data: 54.558 sec elapsed
Step 13: tallying molecules by cell-barcode: 369.919 sec elapsed
Step 14: saving results: 19.665 sec elapsed
Running workflow: 4069.898 sec elapsed

Show workflow output

1. Tranforming data and computing summary statistics

Read counts datatable

data_list$read_counts

Summary statistics of the joined read counts datatable

 data_list$reads_dist_summary$summary_stats

Summary statistics conditional on the PCR duplication level r

 data_list$reads_dist_summary$conditional

The molecular proportions complexity profile

 data_list$pi_r_hat

The molecular proportions complexity profile plot

p_read <- plot_molecules_distributions(data_list, dataset_name, x_lim=120)
plot_grid(p_read$p, 
          p_read$legend,
          ncol=2,
          rel_widths=c(1, 0.1))

Outcome counts datatable

data_list$outcome_counts

Chimera counts datatable

data_list$fit_out$chimera_counts

2. Estimating the sample index hopping rate (SIHR)

GLM fit estimates

data_list$fit_out$glm_estimates

Model fit plot

p_fit <- plot_fit(data_list, dataset_name)

plot_grid(p_fit$p,
          p_fit$legend,
          ncol=2,
          rel_widths=c(1, 0.2))

3. Purging phantom molecules

The optimal cutoff for minimizing the false positive rate

data_list$optimal_cutoff

Plot of the empirical CDF of qs

p_post <-plot_posterior_prob(data_list, dataset_name)
plot_grid(p_post$p,
          p_post$legend,
          ncol=2,
          rel_widths=c(1, 0.1))

Note that q is the marginal posterior distribution of predicted sample of origin s and qs is a tranformation of q.

4. Examining the consequences of index hopping

First we examing the extent of the effects of index hopping on individual samples and then on cell-barcodes.

Tally of predicted phantoms by sample

Here r_a is the number of molecules above the optimal cutoff q, which we predict as real molecules. r_a are the number of molecules below or equal to cutoff and thus we predict as phantoms. f is the number of molecules predicted as phantom no matter what the threshold is. m is the number of total molecules.

data_list$reads_dist_summary$marginal

Tally of predicted phantoms in called cells

The called cells were determined from the unpurged data in order to show the level of contamination by phantom molecules if data were not purged.

data_list$reads_dist_summary$marginal_called_cells

Tally of barcodes by concordance between purged and unpurged data

data_list$called_cells_tally

Table of called cell-barcodes with the highest number of phantoms

data_list$umi_counts_cell %>% 
  map(list("called_cells"))
$P7_0
# A tibble: 1,332 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   475    40   430     5
 2 AGTGAGGTCTGCCCTA   850   432   417     1
 3 CTAACTTAGTTGTAGA   433    34   399     0
 4 CATGCCTTCAATACCG   785   392   391     2
 5 CTCGAGGGTTTCGCTC   396    27   368     1
 6 AGACGTTGTCCGAGTC   403    36   365     2
 7 GAACATCTCCTTGGTC   369    32   336     1
 8 CGGGTCATCCCAAGTA   369    39   329     1
 9 GTCGTAAGTTAAGATG   367    42   325     0
10 AGCATACCATCATCCC   357    37   320     0
# … with 1,322 more rows

$P7_1
# A tibble: 770 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   555    51   500     4
 2 AGACGTTGTCCGAGTC   518    42   471     5
 3 CTAACTTAGTTGTAGA   492    48   443     1
 4 CGGGTCATCCCAAGTA   486    53   432     1
 5 GAATAAGCATATGGTC   464    48   416     0
 6 TGTTCCGAGGCTAGAC   477    70   407     0
 7 GTGGGTCCAAACTGTC   470    87   382     1
 8 CATGCCTTCAATACCG   434    57   377     0
 9 TAAGAGATCTTGGGTA   480   114   366     0
10 GAACATCTCAGAGACG   424    59   365     0
# … with 760 more rows

$P7_10
# A tibble: 2,460 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   293    39   254     0
 2 CTAACTTAGTTGTAGA   255    21   234     0
 3 GAACATCTCCTTGGTC   227    12   214     1
 4 AGACGTTGTCCGAGTC   234    25   209     0
 5 CGGGTCATCCCAAGTA   224    15   209     0
 6 AGTGAGGTCTGCCCTA   218    19   196     3
 7 CTCGAGGGTTTCGCTC   214    18   195     1
 8 GGACATTTCGTAGGAG   207    17   189     1
 9 GTGCTTCGTCCCTACT   205    38   163     4
10 ACATACGCAGCATGAG 12457 12311   103    43
# … with 2,450 more rows

$P7_11
# A tibble: 3,506 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGTGAGGTCTGCCCTA   394    55   338     1
 2 AGGGATGGTCTAACGT   369    30   334     5
 3 CGGGTCATCCCAAGTA   373    62   310     1
 4 CTCGAGGGTTTCGCTC   323    20   302     1
 5 AGACGTTGTCCGAGTC   304    27   274     3
 6 GAACATCTCCTTGGTC   297    22   272     3
 7 CATGCCTTCAATACCG   316    45   270     1
 8 TGTTCCGAGGCTAGAC   284    23   258     3
 9 GGACATTTCGTAGGAG   271    18   247     6
10 GAACATCTCAGAGACG   268    32   232     4
# … with 3,496 more rows

$P7_12
# A tibble: 3,054 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   368    40   322     6
 2 CTAACTTAGTTGTAGA   339    26   313     0
 3 AGTGAGGTCTGCCCTA   331    29   297     5
 4 AGACGTTGTCCGAGTC   303    37   261     5
 5 TGTTCCGAGGCTAGAC   273    21   251     1
 6 GAACATCTCAGAGACG   315    60   250     5
 7 CTCGAGGGTTTCGCTC   271    24   246     1
 8 GAACATCTCCTTGGTC   275    32   242     1
 9 GGCTGGTGTGTCGCTG   258    30   225     3
10 TAAACCGTCTCTGTCG   252    25   224     3
# … with 3,044 more rows

$P7_13
# A tibble: 3,409 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   381    42   336     3
 2 CTAACTTAGTTGTAGA   312    22   286     4
 3 AGTGAGGTCTGCCCTA   308    18   285     5
 4 AGACGTTGTCCGAGTC   290    31   255     4
 5 GAACATCTCCTTGGTC   276    22   251     3
 6 CGGGTCATCCCAAGTA   271    19   246     6
 7 CTCGAGGGTTTCGCTC   260    25   234     1
 8 GAACATCTCAGAGACG   272    43   223     6
 9 TGTTCCGAGGCTAGAC   257    36   218     3
10 GAATAAGCATATGGTC   249    22   216    11
# … with 3,399 more rows

$P7_14
# A tibble: 2,628 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   371    35   334     2
 2 AGACGTTGTCCGAGTC   320    38   282     0
 3 CTAACTTAGTTGTAGA   311    28   282     1
 4 CGGGTCATCCCAAGTA   315    31   281     3
 5 AGTGAGGTCTGCCCTA   301    20   279     2
 6 GAACATCTCCTTGGTC   293    16   277     0
 7 CTCGAGGGTTTCGCTC   268    17   249     2
 8 GAATAAGCATATGGTC   273    43   230     0
 9 GAACATCTCAGAGACG   261    37   224     0
10 TGAGAGGCAACACGCC   239    17   221     1
# … with 2,618 more rows

$P7_15
# A tibble: 2,763 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   322    28   289     5
 2 CTAACTTAGTTGTAGA   242    17   221     4
 3 CGGGTCATCCCAAGTA   234    23   209     2
 4 AGTGAGGTCTGCCCTA   223    19   204     0
 5 AGACGTTGTCCGAGTC   218    18   200     0
 6 GAATAAGCATATGGTC   227    24   192    11
 7 TGAGAGGCAACACGCC   204    14   190     0
 8 GCGCCAAAGGCTATCT   226    80   146     0
 9 CATATGGGTTGCGTTA  1696  1606    89     1
10 CTCGTCATCAGAGGTG  5133  5047    81     5
# … with 2,753 more rows

$P7_2
# A tibble: 4,058 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   347    35   309     3
 2 AGTGAGGTCTGCCCTA   317    26   278    13
 3 CGGGTCATCCCAAGTA   282    28   251     3
 4 GAACATCTCCTTGGTC   268    19   247     2
 5 CTAACTTAGTTGTAGA   266    24   241     1
 6 CTCGAGGGTTTCGCTC   275    36   238     1
 7 GAACATCTCAGAGACG   288    46   238     4
 8 CATGCCTTCAATACCG   277    35   227    15
 9 AGACGTTGTCCGAGTC   253    32   217     4
10 TGTTCCGAGGCTAGAC   232    19   211     2
# … with 4,048 more rows

$P7_3
# A tibble: 3,997 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   376    75   299     2
 2 AGTGAGGTCTGCCCTA   310    30   276     4
 3 CTAACTTAGTTGTAGA   307    37   268     2
 4 AGACGTTGTCCGAGTC   290    23   265     2
 5 GAACATCTCCTTGGTC   275    29   245     1
 6 CTCGAGGGTTTCGCTC   258    22   235     1
 7 TGTTCCGAGGCTAGAC   254    29   223     2
 8 CGGGTCATCCCAAGTA   249    30   218     1
 9 CATCCACAGATGGCGT   238    22   215     1
10 CATGCCTTCAATACCG   253    31   213     9
# … with 3,987 more rows

$P7_4
# A tibble: 996 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 GCTGCTTGTCCGAAGA   764   158   606     0
 2 GTGCTTCGTCCCTACT   570    86   483     1
 3 AGGGATGGTCTAACGT   396    41   352     3
 4 TGGGAAGCATTCGACA   358    18   340     0
 5 CTAACTTAGTTGTAGA   324    25   297     2
 6 AGTGAGGTCTGCCCTA   329    33   295     1
 7 AGGCCGTGTGCGATAG   309    16   293     0
 8 AGACGTTGTCCGAGTC   295    26   268     1
 9 AGCGGTCTCTTAGCCC   285    20   265     0
10 CAGATCAAGACAGAGA   289    27   262     0
# … with 986 more rows

$P7_5
# A tibble: 1,979 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   272    27   243     2
 2 CGGGTCATCCCAAGTA   251    26   224     1
 3 CTAACTTAGTTGTAGA   241    22   219     0
 4 AGTGAGGTCTGCCCTA 12662 12418   210    34
 5 GAACATCTCCTTGGTC   224    21   203     0
 6 CTCGAGGGTTTCGCTC   216    15   201     0
 7 AGACGTTGTCCGAGTC   217    18   197     2
 8 TAAACCGTCTCTGTCG   213    20   193     0
 9 GAATAAGCATATGGTC   222    29   183    10
10 GGACATTTCGTAGGAG   257    73   182     2
# … with 1,969 more rows

$P7_6
# A tibble: 7,291 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGTGAGGTCTGCCCTA   266    16   248     2
 2 AGGGATGGTCTAACGT   248    31   217     0
 3 GAACATCTCCTTGGTC   238    22   216     0
 4 CGGGTCATCCCAAGTA   210    19   191     0
 5 CTCGAGGGTTTCGCTC   208    15   191     2
 6 GTGCTTCGTCCCTACT   219    32   182     5
 7 GAATAAGCATATGGTC   208    17   181    10
 8 CTAACTTAGTTGTAGA   201    22   178     1
 9 AAATGCCTCCCAAGTA   208    29   176     3
10 TAAGAGATCTTGGGTA  5376  5228   126    22
# … with 7,281 more rows

$P7_7
# A tibble: 742 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   301    30   269     2
 2 AGTGAGGTCTGCCCTA   258    20   238     0
 3 CTAACTTAGTTGTAGA   247    10   235     2
 4 CTCGAGGGTTTCGCTC   234    17   216     1
 5 GAACATCTCAGAGACG   246    33   213     0
 6 TAAACCGTCTCTGTCG   228    14   211     3
 7 GAACATCTCCTTGGTC   235    25   209     1
 8 AGACGTTGTCCGAGTC   222    17   204     1
 9 GGACATTTCGTAGGAG   219    28   191     0
10 TAAACCGCATGTAGTC   212    22   187     3
# … with 732 more rows

$P7_8
# A tibble: 1,199 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGTGAGGTCTGCCCTA   505     1   465    39
 2 GAATAAGCATATGGTC   468     1   420    47
 3 GAACATCTCCTTGGTC   442     2   408    32
 4 CTCGAGGGTTTCGCTC   410     0   389    21
 5 TGTTCCGAGGCTAGAC   421     0   377    44
 6 GAACATCTCAGAGACG   415     2   363    50
 7 GAACGGAGTTGCGCAC   406     5   361    40
 8 CATGCCTTCAATACCG   405     1   356    48
 9 GTGCTTCGTCCCTACT   425     1   344    80
10 AAATGCCTCCCAAGTA   391     2   341    48
# … with 1,189 more rows

$P7_9
# A tibble: 2,396 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 AGGGATGGTCTAACGT   446    42   403     1
 2 AGTGAGGTCTGCCCTA   409    31   376     2
 3 CGGGTCATCCCAAGTA   411    34   374     3
 4 CTAACTTAGTTGTAGA   385    27   356     2
 5 CTCGAGGGTTTCGCTC   375    26   346     3
 6 AGACGTTGTCCGAGTC   365    26   338     1
 7 GAACATCTCCTTGGTC   359    29   330     0
 8 GAATAAGCATATGGTC   337    34   303     0
 9 GGCTGGTGTGTCGCTG   316    18   292     6
10 GAACATCTCAGAGACG   329    45   284     0
# … with 2,386 more rows

Table of background cell-barcodes with highest number of phantoms

data_list$umi_counts_cell %>% 
  map(list("background_cells"))
$P7_0
# A tibble: 307,833 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   449    30   413     6
 2 GCTGCTTGTCCGAAGA   533   222   302     9
 3 AAATGCCTCCCAAGTA   353    54   296     3
 4 GGACATTTCGTAGGAG   315    34   277     4
 5 TAAACCGCATGTAGTC   278    33   238     7
 6 GCAAACTAGCTTATCG   253    13   237     3
 7 ACACCAACATAAGACA   238     8   227     3
 8 ACGCCGAGTCTACCTC   234    22   210     2
 9 TTAACTCCAATGGTCT   228    18   209     1
10 AACTCAGTCTAACGGT   226    16   208     2
# … with 307,823 more rows

$P7_1
# A tibble: 305,606 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 GCTGCTTGTCCGAAGA   546   145   400     1
 2 AGAGTGGTCATGTCCC   199    10   188     1
 3 GACCAATAGTACGACG   200    14   186     0
 4 CTCGTCATCAGAGGTG   200    14   185     1
 5 TGGCTGGAGAAGGTGA   196    12   184     0
 6 AGGGATGAGTACTTGC   197    15   182     0
 7 GCTTCCAGTTCCGGCA   195    11   182     2
 8 CATGGCGGTGCGATAG   198    17   180     1
 9 CCACCTAAGGCCCGTT   199    19   180     0
10 GGCAATTTCTTCTGGC   195    15   180     0
# … with 305,596 more rows

$P7_10
# A tibble: 417,050 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   225    17   205     3
 2 GAATAAGCATATGGTC   227    34   188     5
 3 GATCTAGTCGAGAACG   189    15   174     0
 4 TGTTCCGAGGCTAGAC   195    19   171     5
 5 GAACATCTCAGAGACG   197    31   166     0
 6 GGCTGGTGTGTCGCTG   182    14   165     3
 7 TAAGAGATCTTGGGTA   196    26   164     6
 8 GCTGCTTGTCCGAAGA   243    73   163     7
 9 AAATGCCTCCCAAGTA   183    20   162     1
10 GAACGGAGTTGCGCAC   190    19   162     9
# … with 417,040 more rows

$P7_11
# A tibble: 323,694 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CTAACTTAGTTGTAGA   332    26   304     2
 2 CACACCTTCAACGCTA   323    23   298     2
 3 GCTGCTTGTCCGAAGA   403   131   263     9
 4 TAAACCGTCTCTGTCG   272    21   250     1
 5 GAATAAGCATATGGTC   290    24   249    17
 6 AAATGCCTCCCAAGTA   288    36   248     4
 7 CATCCACAGATGGCGT   243    20   218     5
 8 GATCTAGTCGAGAACG   236    19   217     0
 9 GAACGGAGTTGCGCAC   263    39   213    11
10 GCAAACTTCGACAGCC   233    18   213     2
# … with 323,684 more rows

$P7_12
# A tibble: 351,076 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CGGGTCATCCCAAGTA   342    37   302     3
 2 CACACCTTCAACGCTA   292    31   261     0
 3 GCTGCTTGTCCGAAGA   383   110   259    14
 4 GAATAAGCATATGGTC   300    34   255    11
 5 GAACGGAGTTGCGCAC   276    32   232    12
 6 ACTTACTCAGCTCGAC   225    16   206     3
 7 TAAGAGATCTTGGGTA   239    32   202     5
 8 GCAAACTTCGACAGCC   230    29   198     3
 9 AAACGGGAGGATATAC   219    26   191     2
10 GGATGTTAGGGAACGG   204    11   191     2
# … with 351,066 more rows

$P7_13
# A tibble: 358,946 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   275    25   245     5
 2 GCTGCTTGTCCGAAGA   311    91   213     7
 3 CGGACACAGCGCTTAT   223    15   206     2
 4 CATCCACAGATGGCGT   215    29   184     2
 5 GAACGGAGTTGCGCAC   225    36   183     6
 6 GCTGCGACAGTAAGCG   190     8   181     1
 7 GGCTGGTGTGTCGCTG   198    16   178     4
 8 ACATACGCAGCATGAG   198    24   174     0
 9 ACTTTCAGTTTGACTG   194    18   174     2
10 CGACTTCAGACCTAGG   194    20   174     0
# … with 358,936 more rows

$P7_14
# A tibble: 308,825 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   306    21   284     1
 2 GAACGGAGTTGCGCAC   257    25   232     0
 3 GCTGCTTGTCCGAAGA   309    77   225     7
 4 GATCTAGTCGAGAACG   240    32   208     0
 5 CATCCACAGATGGCGT   223    21   199     3
 6 GCAAACTTCGACAGCC   213    21   190     2
 7 ACTTACTCAGCTCGAC   206    16   189     1
 8 AAATGCCGTGAACCTT   197    13   184     0
 9 CGACTTCAGACCTAGG   198    15   182     1
10 TGGGAAGTCAACCAAC   198    17   181     0
# … with 308,815 more rows

$P7_15
# A tibble: 296,794 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 TAAACCGTCTCTGTCG   198    11   187     0
 2 GAACGGAGTTGCGCAC   212    24   180     8
 3 TGTTCCGAGGCTAGAC   195    22   173     0
 4 CACACCTTCAACGCTA   193    18   172     3
 5 CTCGAGGGTTTCGCTC   188    18   170     0
 6 GAACATCTCCTTGGTC   180    10   170     0
 7 CATGCCTTCAATACCG   197    31   166     0
 8 GGCTGGTGTGTCGCTG   182    16   166     0
 9 GTGCTTCGTCCCTACT   201    33   165     3
10 ACTTACTCAGCTCGAC   177    13   162     2
# … with 296,784 more rows

$P7_2
# A tibble: 360,298 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   254    18   234     2
 2 GATCTAGTCGAGAACG   219    16   201     2
 3 GAATAAGCATATGGTC   238    32   196    10
 4 ACTTACTCAGCTCGAC   228    35   192     1
 5 GCTGCTTGTCCGAAGA   271    69   191    11
 6 CGGACACAGCGCTTAT   208    25   181     2
 7 GAACGGAGTTGCGCAC   208    20   181     7
 8 GCTGCGACAGTAAGCG   197    17   180     0
 9 AAACGGGAGGATATAC   189    11   176     2
10 CATCCACAGATGGCGT   197    25   170     2
# … with 360,288 more rows

$P7_3
# A tibble: 379,886 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   247    18   223     6
 2 GCTGCTTGTCCGAAGA   306    84   215     7
 3 GAATAAGCATATGGTC   263    43   209    11
 4 TAAACCGTCTCTGTCG   231    20   205     6
 5 AGATTGCAGAGCAATT   200    13   187     0
 6 ACTTTCAGTTTGACTG   198    19   179     0
 7 TCGAGGCTCCCTCTTT   194    15   178     1
 8 GAACGGAGTTGCGCAC   214    29   176     9
 9 CTCAGAACATATGCTG   194    19   170     5
10 CGGACACAGCGCTTAT   188    19   168     1
# … with 379,876 more rows

$P7_4
# A tibble: 314,515 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   320    16   304     0
 2 CGTCTACTCTTGACGA   195     9   186     0
 3 CCTACCACATCGGTTA   197    19   178     0
 4 GACTACAGTTCAACCA   189    11   178     0
 5 GGTGAAGAGGTGATAT   189    12   177     0
 6 ACTTTCAGTTTGACTG   192    14   176     2
 7 AACGTTGTCAACACAC   187    11   175     1
 8 AAATGCCGTGAACCTT   199    25   174     0
 9 TCGAGGCTCGCGCCAA   191    17   174     0
10 TGTTCCGCACTTAAGC   187    13   174     0
# … with 314,505 more rows

$P7_5
# A tibble: 363,946 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   287    40   244     3
 2 GCTGCTTGTCCGAAGA   301    79   217     5
 3 TGAGAGGCAACACGCC   198    15   180     3
 4 AAATGCCTCCCAAGTA   195    22   169     4
 5 TGTTCCGAGGCTAGAC   191    21   167     3
 6 ACTTACTCAGCTCGAC   181    13   166     2
 7 AGATTGCAGAGCAATT   167     8   159     0
 8 CCGTACTGTCAGATAA   178    17   159     2
 9 TAAGAGATCTTGGGTA   182    20   156     6
10 TAAACCGCATGTAGTC   174    17   154     3
# … with 363,936 more rows

$P7_6
# A tibble: 349,960 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   220    11   204     5
 2 GCTGCTTGTCCGAAGA   268    78   187     3
 3 AGACGTTGTCCGAGTC   195    15   180     0
 4 TGTTCCGAGGCTAGAC   195    21   170     4
 5 GGACATTTCGTAGGAG   187    14   168     5
 6 TGAGAGGCAACACGCC   179    16   162     1
 7 AAATGCCGTGAACCTT   167    11   155     1
 8 TCTATTGCATAAAGGT   168    13   155     0
 9 ACTTACTCAGCTCGAC   169    15   153     1
10 CATGCCTTCAATACCG   182    22   152     8
# … with 349,950 more rows

$P7_7
# A tibble: 273,890 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   230    10   219     1
 2 CGGGTCATCCCAAGTA   235    25   210     0
 3 CATCCACAGATGGCGT   196    12   184     0
 4 GTGCTTCGTCCCTACT   235    48   184     3
 5 TGTTCCGAGGCTAGAC   199    21   178     0
 6 GCTGCTTGTCCGAAGA   253    75   173     5
 7 GATCTAGTCGAGAACG   190    20   170     0
 8 GCTGCGACAGTAAGCG   179     9   169     1
 9 AAATGCCTCCCAAGTA   199    30   168     1
10 ACTTACTCAGCTCGAC   181    13   168     0
# … with 273,880 more rows

$P7_8
# A tibble: 270,526 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   470     0   432    38
 2 GCTGCTTGTCCGAAGA   473     6   336   131
 3 GCAAACTAGCTTATCG   299     0   284    15
 4 GTAGGCCAGCCGATTT   309     0   270    39
 5 ACACCAACATAAGACA   247     1   232    14
 6 GCTTCCACACTTCGAA   251     0   223    28
 7 TTAACTCCAATGGTCT   252     7   222    23
 8 GCACTCTAGATACACA   243     0   218    25
 9 ACGCCGAGTCTACCTC   233     0   215    18
10 CTTAACTCACTACAGT   224     0   210    14
# … with 270,516 more rows

$P7_9
# A tibble: 314,581 x 5
   cell                 m   r_a     f   r_b
   <chr>            <int> <dbl> <int> <dbl>
 1 CACACCTTCAACGCTA   389    27   358     4
 2 GAACGGAGTTGCGCAC   362    44   318     0
 3 AAATGCCTCCCAAGTA   322    29   292     1
 4 ACTTACTCAGCTCGAC   290    33   256     1
 5 CATCCACAGATGGCGT   274    24   249     1
 6 GTAGGCCAGCCGATTT   272    37   235     0
 7 GATCTAGTCGAGAACG   254    19   233     2
 8 CGGACACAGCGCTTAT   259    28   230     1
 9 AAACGGGAGGATATAC   264    35   228     1
10 GGATGTTAGGGAACGG   247    27   219     1
# … with 314,571 more rows

Session Info

# memory usage
gc()
             used    (Mb)  gc trigger     (Mb)    max used     (Mb)
Ncells    6356772   339.5    10309335    550.6    10309335    550.6
Vcells 5305151352 40475.1 19670404828 150073.3 30723203766 234399.5
sessionInfo()
R version 3.5.2 (2018-12-20)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

Matrix products: default
BLAS/LAPACK: /cvmfs/soft.computecanada.ca/easybuild/software/2017/Core/imkl/2018.3.222/compilers_and_libraries_2018.3.222/linux/mkl/lib/intel64_lin/libmkl_gf_lp64.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [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] rhdf5_2.26.2       cowplot_0.9.4      data.table_1.12.0 
 [4] tictoc_1.0         furrr_0.1.0        future_1.11.1.1   
 [7] broom_0.5.1        matrixStats_0.54.0 forcats_0.4.0     
[10] stringr_1.4.0      dplyr_0.8.0.1      purrr_0.3.1       
[13] readr_1.3.1        tidyr_0.8.3        tibble_2.0.1      
[16] ggplot2_3.1.0      tidyverse_1.2.1    rmarkdown_1.11    

loaded via a namespace (and not attached):
 [1] nlme_3.1-137                bitops_1.0-6               
 [3] lubridate_1.7.4             httr_1.4.0                 
 [5] rprojroot_1.3-2             GenomeInfoDb_1.18.2        
 [7] tools_3.5.2                 backports_1.1.3            
 [9] utf8_1.1.4                  R6_2.4.0                   
[11] HDF5Array_1.10.1            lazyeval_0.2.1             
[13] BiocGenerics_0.28.0         colorspace_1.4-0           
[15] withr_2.1.2                 tidyselect_0.2.5           
[17] compiler_3.5.2              cli_1.0.1                  
[19] rvest_0.3.2                 Biobase_2.42.0             
[21] xml2_1.2.0                  DelayedArray_0.8.0         
[23] labeling_0.3                scales_1.0.0               
[25] digest_0.6.18               XVector_0.22.0             
[27] base64enc_0.1-3             pkgconfig_2.0.2            
[29] htmltools_0.3.6             limma_3.38.3               
[31] rlang_0.3.1                 readxl_1.3.0               
[33] rstudioapi_0.9.0            generics_0.0.2             
[35] jsonlite_1.6                BiocParallel_1.16.6        
[37] RCurl_1.95-4.12             magrittr_1.5               
[39] GenomeInfoDbData_1.2.0      Matrix_1.2-15              
[41] Rcpp_1.0.0                  munsell_0.5.0              
[43] S4Vectors_0.20.1            Rhdf5lib_1.4.2             
[45] fansi_0.4.0                 stringi_1.3.1              
[47] yaml_2.2.0                  edgeR_3.24.3               
[49] MASS_7.3-51.1               SummarizedExperiment_1.12.0
[51] zlibbioc_1.28.0             plyr_1.8.4                 
[53] grid_3.5.2                  parallel_3.5.2             
[55] listenv_0.7.0               crayon_1.3.4               
[57] lattice_0.20-38             haven_2.1.0                
[59] hms_0.4.2                   locfit_1.5-9.1             
[61] knitr_1.21                  pillar_1.3.1               
[63] GenomicRanges_1.34.0        codetools_0.2-16           
[65] stats4_3.5.2                glue_1.3.0                 
[67] evaluate_0.13               modelr_0.1.4               
[69] cellranger_1.1.0            gtable_0.2.0               
[71] assertthat_0.2.0            xfun_0.5                   
[73] DropletUtils_1.2.2          viridisLite_0.3.0          
[75] SingleCellExperiment_1.4.1  IRanges_2.16.0             
[77] globals_0.12.4
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