1 PBMC Infection Differential Expression, Infection: 20180822 Rundate: 20181206

This document turns to the infection of PBMC cells with L.panamensis. This data is particularly strangely affected by the different strains used to infect the cells, and as a result is both useful and troubling.

Given the observations above, we have some ideas of ways to pass the data for differential expression analyses which may or may not be ‘better’. Lets try some and see what happens.

1.1 Create data sets to compare differential expression analyses

Given the above ways to massage the data, lets use a few of them for limma/deseq/edger. The main caveat in this is that those tools really do expect specific distributions of data which we horribly violate if we use log2() data, which is why in the previous blocks I named them l2blahblah, thus we can do the same sets of normalization but without that and forcibly push the resulting data into limma/edger/deseq.

2 The negative control

Everything I did in 02_estimation_infection.html suggests that there are no significant differences visible if one looks just at chronic/self-healing in this data. Further testing has seemingly proven this statement, as a result most of the following analyses will look at chronic/uninfected and self-healing/uninfected followed by attempts to reconcile those results.

2.1 Filter the data

To save some time and annoyance with sva, lets filter the data now. In addition, write down a small function used to extract the sets of significant genes across different contrasts (notably self/uninfected vs. chronic/uninfected).

hs_inf_filt <- sm(normalize_expt(hs_cds_inf, filter=TRUE))
hs_uninf_filt <- sm(normalize_expt(hs_cds_uninf, filter=TRUE))
keepers <- list("sh_nil" = c("sh", "uninf"),
                "ch_nil" = c("chr", "uninf"),
                "ch_sh" = c("chr", "sh"))

3 Queries from 20181205

3.1 Each individual, chr vs sh.

d107 <- subset_expt(hs_inf_filt, subset="donor=='d107'")

## There were 18, now there are 6 samples.

d107_pairwise <- all_pairwise(d107, model_batch=FALSE)

## Assuming no batch in model for testing pca.

## Finished running DE analyses, collecting outputs.

## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr

d107_table <- combine_de_tables(d107_pairwise)

## Writing a legend of columns.

## Working on table 1/1: sh_vs_chr

summary(d107_table$data)

##           Length Class      Mode
## sh_vs_chr 46     data.frame list

d107_sig <- extract_significant_genes(d107_table, according_to="deseq", p=0.1, p_type="raw")

## Writing a legend of columns.

## The count is: 1 and the test is: deseq.

## Writing excel data according to deseq for sh_vs_chr: 1/1.

## After (adj)p filter, the up genes table has 197 genes.

## After (adj)p filter, the down genes table has 403 genes.

## After fold change filter, the up genes table has 5 genes.

## After fold change filter, the down genes table has 4 genes.

## Printing significant genes to the file: excel/significant_genes.xlsx

## 1/1: Creating significant table up_1deseq_sh_vs_chr

## Adding significance bar plots.

head(d107_sig$deseq$ups[[1]])

##                 ensembltranscriptid   ensemblgeneid hgncsymbol
## ENSG00000205846     ENST00000382073 ENSG00000205846     CLEC6A
## ENSG00000111052     ENST00000261203 ENSG00000111052      LIN7A
## ENSG00000184838     ENST00000379551 ENSG00000184838      PRR16
## ENSG00000186047     ENST00000400393 ENSG00000186047      DLEU7
## ENSG00000125538     ENST00000263341 ENSG00000125538       IL1B
##                                                                                                 description
## ENSG00000205846                  C-type lectin domain family 6 member A [Source:HGNC Symbol;Acc:HGNC:14556]
## ENSG00000111052 lin-7 homolog A, crumbs cell polarity complex component [Source:HGNC Symbol;Acc:HGNC:17787]
## ENSG00000184838                                         proline rich 16 [Source:HGNC Symbol;Acc:HGNC:29654]
## ENSG00000186047                      deleted in lymphocytic leukemia, 7 [Source:HGNC Symbol;Acc:HGNC:17567]
## ENSG00000125538                                       interleukin 1 beta [Source:HGNC Symbol;Acc:HGNC:5992]
##                    genebiotype deseq_logfc deseq_adjp edger_logfc
## ENSG00000205846 protein_coding       1.554     0.9995       1.545
## ENSG00000111052 protein_coding       1.242     0.9995       1.194
## ENSG00000184838 protein_coding       1.215     0.9995       1.204
## ENSG00000186047 protein_coding       1.134     0.9995       1.155
## ENSG00000125538 protein_coding       1.122     0.9995       1.128
##                 edger_adjp limma_logfc limma_adjp basic_nummed
## ENSG00000205846     0.9996      1.6870     0.9842       1.1460
## ENSG00000111052     0.9996      1.3100     0.9842      -0.3794
## ENSG00000184838     0.9996      1.3110     0.9842       1.3110
## ENSG00000186047     0.9996      1.3360     0.9842       0.3444
## ENSG00000125538     0.9996      0.9156     0.9842       6.6270
##                 basic_denmed basic_numvar basic_denvar basic_logfc basic_t
## ENSG00000205846     -0.61650    3.218e-01    6.981e-01      1.7630   2.866
## ENSG00000111052     -1.10200    4.981e-01    8.553e-01      0.7225   2.042
## ENSG00000184838      0.03563    2.780e-02    5.732e-01      1.2750   3.031
## ENSG00000186047     -1.25400    4.012e-02    3.564e-01      1.5980   3.524
## ENSG00000125538      4.94900    1.440e+00    1.499e-01      1.6780   1.227
##                   basic_p basic_adjp deseq_basemean deseq_lfcse deseq_stat
## ENSG00000205846 5.311e-02  9.994e-01         17.520      0.5644      2.754
## ENSG00000111052 1.156e-01  9.994e-01          9.367      0.7201      1.724
## ENSG00000184838 8.370e-02  9.994e-01         22.120      0.4650      2.613
## ENSG00000186047 5.337e-02  9.994e-01         11.360      0.6167      1.839
## ENSG00000125538 3.261e-01  9.994e-01        750.200      0.6455      1.738
##                  deseq_p ebseq_fc ebseq_logfc ebseq_postfc ebseq_mean
## ENSG00000205846 0.005894    2.749      1.4590        2.638     16.996
## ENSG00000111052 0.084630    2.240      1.1634        2.121      9.468
## ENSG00000184838 0.008967    2.193      1.1329        2.141     21.699
## ENSG00000186047 0.065850    1.919      0.9400        1.850     11.249
## ENSG00000125538 0.082280    1.942      0.9572        1.940    725.275
##                 ebseq_ppee ebseq_ppde ebseq_adjp edger_logcpm edger_lr
## ENSG00000205846     0.4349  5.651e-01     0.4349      0.71380   12.460
## ENSG00000111052     0.9723  2.771e-02     0.9723     -0.26660    4.806
## ENSG00000184838     1.0000  7.535e-18     1.0000      0.96760   11.230
## ENSG00000186047     1.0000  1.933e-19     1.0000      0.08541    6.563
## ENSG00000125538     0.9905  9.522e-03     0.9905      5.84800    8.959
##                   edger_p limma_ave limma_t limma_b  limma_p
## ENSG00000205846 0.0004150   0.06708   3.668  -4.412 0.004893
## ENSG00000111052 0.0283500  -0.76930   2.413  -4.518 0.038260
## ENSG00000184838 0.0008061   0.52320   3.822  -4.359 0.003837
## ENSG00000186047 0.0104100  -0.42420   3.653  -4.441 0.005006
## ENSG00000125538 0.0027600   5.47900   1.804  -4.258 0.103700
##                 limma_adjp_fdr deseq_adjp_fdr edger_adjp_fdr
## ENSG00000205846      9.842e-01      9.997e-01      9.995e-01
## ENSG00000111052      9.842e-01      9.997e-01      9.995e-01
## ENSG00000184838      9.842e-01      9.997e-01      9.995e-01
## ENSG00000186047      9.842e-01      9.997e-01      9.995e-01
## ENSG00000125538      9.842e-01      9.997e-01      9.995e-01
##                 basic_adjp_fdr lfc_meta   lfc_var lfc_varbymed    p_meta
## ENSG00000205846      9.993e-01    1.595 0.000e+00    0.000e+00 3.734e-03
## ENSG00000111052      9.993e-01    1.237 1.337e-03    1.081e-03 5.041e-02
## ENSG00000184838      9.993e-01    1.245 8.533e-05    6.852e-05 4.537e-03
## ENSG00000186047      9.993e-01    1.218 1.058e-03    8.682e-04 2.709e-02
## ENSG00000125538      9.993e-01    1.061 3.106e-02    2.928e-02 6.291e-02
##                     p_var
## ENSG00000205846 8.512e-06
## ENSG00000111052 9.026e-04
## ENSG00000184838 1.702e-05
## ENSG00000186047 1.134e-03
## ENSG00000125538 2.829e-03

d107_ma <- extract_de_plots(d107_pairwise, p=0.1, p_type="raw")$ma$plot
d107_ma

plot_histogram(d107_table[["data"]][["sh_vs_chr"]][, c("deseq_p")])

d108 <- subset_expt(hs_inf_filt, subset="donor=='d108'")

## There were 18, now there are 6 samples.

d108_pairwise <- all_pairwise(d108, model_batch=FALSE)

## Assuming no batch in model for testing pca.

## Finished running DE analyses, collecting outputs.

## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr

d108_table <- combine_de_tables(d108_pairwise)

## Writing a legend of columns.

## Working on table 1/1: sh_vs_chr

summary(d108_table$data)

##           Length Class      Mode
## sh_vs_chr 46     data.frame list

d108_sig <- extract_significant_genes(d108_table, according_to="deseq", p=0.1, p_type="raw")

## Writing a legend of columns.

## The count is: 1 and the test is: deseq.

## Writing excel data according to deseq for sh_vs_chr: 1/1.

## After (adj)p filter, the up genes table has 91 genes.

## After (adj)p filter, the down genes table has 129 genes.

## After fold change filter, the up genes table has 2 genes.

## After fold change filter, the down genes table has 3 genes.

## Printing significant genes to the file: excel/significant_genes.xlsx

## 1/1: Creating significant table up_1deseq_sh_vs_chr

## Adding significance bar plots.

head(d108_sig$deseq$ups[[1]])

##                 ensembltranscriptid   ensemblgeneid hgncsymbol
## ENSG00000119457     ENST00000374228 ENSG00000119457    SLC46A2
## ENSG00000162493     ENST00000294489 ENSG00000162493       PDPN
##                                                                           description
## ENSG00000119457 solute carrier family 46 member 2 [Source:HGNC Symbol;Acc:HGNC:16055]
## ENSG00000162493                        podoplanin [Source:HGNC Symbol;Acc:HGNC:29602]
##                    genebiotype deseq_logfc deseq_adjp edger_logfc
## ENSG00000119457 protein_coding       1.124     0.9999       1.110
## ENSG00000162493 protein_coding       1.054     0.9999       1.047
##                 edger_adjp limma_logfc limma_adjp basic_nummed
## ENSG00000119457          1      1.0310     0.9984       0.7973
## ENSG00000162493          1      0.9181     0.9984       1.5110
##                 basic_denmed basic_numvar basic_denvar basic_logfc basic_t
## ENSG00000119457     -0.09023    2.498e-01    5.520e-02      0.8875   3.286
## ENSG00000162493      0.79730    9.253e-01    5.631e-01      0.7134   1.414
##                   basic_p basic_adjp deseq_basemean deseq_lfcse deseq_stat
## ENSG00000119457 4.996e-02      1e+00          14.09      0.5510      2.039
## ENSG00000162493 2.343e-01      1e+00          22.92      0.5347      1.971
##                 deseq_p ebseq_fc ebseq_logfc ebseq_postfc ebseq_mean
## ENSG00000119457 0.04144    2.246       1.167        2.187      14.19
## ENSG00000162493 0.04873    2.177       1.122        2.143      23.12
##                 ebseq_ppee ebseq_ppde ebseq_adjp edger_logcpm edger_lr
## ENSG00000119457          1  2.896e-05          1       0.5923    6.374
## ENSG00000162493          1  4.423e-06          1       1.2200    4.785
##                 edger_p limma_ave limma_t limma_b limma_p limma_adjp_fdr
## ENSG00000119457 0.01158    0.2681   2.918  -4.453 0.01789      9.984e-01
## ENSG00000162493 0.02871    0.8417   1.655  -4.532 0.13370      9.984e-01
##                 deseq_adjp_fdr edger_adjp_fdr basic_adjp_fdr lfc_meta
## ENSG00000119457      1.000e+00      1.000e+00          1e+00    1.088
## ENSG00000162493      1.000e+00      1.000e+00          1e+00    1.003
##                   lfc_var lfc_varbymed    p_meta     p_var
## ENSG00000119457 0.000e+00    0.000e+00 2.364e-02 2.477e-04
## ENSG00000162493 3.053e-03    3.043e-03 7.038e-02 3.107e-03

d108_ma <- extract_de_plots(d108_pairwise, p=0.1, p_type="raw")$ma$plot
d108_ma

plot_histogram(d108_table[["data"]][["sh_vs_chr"]][, c("deseq_p")])

d110 <- subset_expt(hs_inf_filt, subset="donor=='d110'")

## There were 18, now there are 6 samples.

d110_pairwise <- all_pairwise(d110, model_batch=FALSE)

## Assuming no batch in model for testing pca.

## Finished running DE analyses, collecting outputs.

## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr
## Comparing analyses 1/1: sh_vs_chr

d110_table <- combine_de_tables(d110_pairwise)

## Writing a legend of columns.

## Working on table 1/1: sh_vs_chr

summary(d110_table$data)

##           Length Class      Mode
## sh_vs_chr 46     data.frame list

d110_sig <- extract_significant_genes(d110_table, according_to="deseq", p=0.1, p_type="raw")

## Writing a legend of columns.

## The count is: 1 and the test is: deseq.

## Writing excel data according to deseq for sh_vs_chr: 1/1.

## After (adj)p filter, the up genes table has 1017 genes.

## After (adj)p filter, the down genes table has 1472 genes.

## After fold change filter, the up genes table has 6 genes.

## After fold change filter, the down genes table has 35 genes.

## Printing significant genes to the file: excel/significant_genes.xlsx

## 1/1: Creating significant table up_1deseq_sh_vs_chr

## Adding significance bar plots.

head(d110_sig$deseq$ups[[1]])

##                 ensembltranscriptid   ensemblgeneid hgncsymbol
## ENSG00000143333     ENST00000367558 ENSG00000143333      RGS16
## ENSG00000120436     ENST00000366834 ENSG00000120436      GPR31
## ENSG00000099985     ENST00000215781 ENSG00000099985        OSM
## ENSG00000137834     ENST00000288840 ENSG00000137834      SMAD6
## ENSG00000177807     ENST00000368089 ENSG00000177807     KCNJ10
## ENSG00000114115     ENST00000232219 ENSG00000114115       RBP1
##                                                                                              description
## ENSG00000143333                   regulator of G-protein signaling 16 [Source:HGNC Symbol;Acc:HGNC:9997]
## ENSG00000120436                         G protein-coupled receptor 31 [Source:HGNC Symbol;Acc:HGNC:4486]
## ENSG00000099985                                          oncostatin M [Source:HGNC Symbol;Acc:HGNC:8506]
## ENSG00000137834                                  SMAD family member 6 [Source:HGNC Symbol;Acc:HGNC:6772]
## ENSG00000177807 potassium voltage-gated channel subfamily J member 10 [Source:HGNC Symbol;Acc:HGNC:6256]
## ENSG00000114115                             retinol binding protein 1 [Source:HGNC Symbol;Acc:HGNC:9919]
##                    genebiotype deseq_logfc deseq_adjp edger_logfc
## ENSG00000143333 protein_coding       1.510  6.358e-05      1.4950
## ENSG00000120436 protein_coding       1.478  1.000e+00      1.4550
## ENSG00000099985 protein_coding       1.226  5.618e-06      1.2100
## ENSG00000137834 protein_coding       1.155  1.000e+00      1.1340
## ENSG00000177807 protein_coding       1.035  1.593e-01      1.0190
## ENSG00000114115 protein_coding       1.012  1.000e+00      0.9872
##                 edger_adjp limma_logfc limma_adjp basic_nummed
## ENSG00000143333  4.147e-06      1.5450     0.2282       2.7890
## ENSG00000120436  5.293e-02      1.4270     0.2438       1.1780
## ENSG00000099985  1.314e-06      1.2660     0.2282       5.6910
## ENSG00000137834  8.498e-02      1.1040     0.2282       0.7412
## ENSG00000177807  1.213e-01      0.9465     0.3169       3.0750
## ENSG00000114115  1.662e-01      1.1950     0.2322       1.0880
##                 basic_denmed basic_numvar basic_denvar basic_logfc basic_t
## ENSG00000143333      1.47200    1.027e-01    1.040e-01       1.317   5.797
## ENSG00000120436     -0.78610    8.451e-01    1.678e-01       1.964   2.381
## ENSG00000099985      4.19200    6.852e-02    1.715e-01       1.499   4.497
## ENSG00000137834     -0.55160    1.598e-01    8.963e-02       1.293   3.828
## ENSG00000177807      1.91700    9.142e-01    1.241e-01       1.158   1.502
## ENSG00000114115     -0.04912    1.870e-02    8.025e-01       1.137   2.294
##                   basic_p basic_adjp deseq_basemean deseq_lfcse deseq_stat
## ENSG00000143333 4.404e-03  4.805e-01          68.76      0.2758      5.477
## ENSG00000120436 1.049e-01  4.810e-01          16.89      0.5436      2.719
## ENSG00000099985 1.587e-02  4.805e-01         484.20      0.2032      6.031
## ENSG00000137834 2.146e-02  4.805e-01          15.79      0.4878      2.368
## ENSG00000177807 2.461e-01  5.597e-01          73.35      0.3745      2.765
## ENSG00000114115 1.432e-01  4.929e-01          21.96      0.4624      2.188
##                   deseq_p ebseq_fc ebseq_logfc ebseq_postfc ebseq_mean
## ENSG00000143333 4.315e-08    2.874       1.523        2.851      69.31
## ENSG00000120436 6.544e-03    2.807       1.489        2.719      16.93
## ENSG00000099985 1.634e-09    2.396       1.261        2.394     488.89
## ENSG00000137834 1.787e-02    2.256       1.174        2.201      15.95
## ENSG00000177807 5.700e-03    2.060       1.043        2.050      73.55
## ENSG00000114115 2.866e-02    2.073       1.051        2.040      22.14
##                 ebseq_ppee ebseq_ppde ebseq_adjp edger_logcpm edger_lr
## ENSG00000143333  4.467e-10  1.000e+00  4.467e-10       2.2710   36.250
## ENSG00000120436  2.182e-01  7.818e-01  2.182e-01       0.3850   11.210
## ENSG00000099985  1.874e-06  1.000e+00  1.874e-06       5.0590   39.490
## ENSG00000137834  1.000e+00  1.816e-12  1.000e+00       0.2622    9.693
## ENSG00000177807  7.014e-01  2.986e-01  7.014e-01       2.3790    8.339
## ENSG00000114115  2.122e-01  7.878e-01  2.122e-01       0.7058    6.987
##                   edger_p limma_ave limma_t limma_b   limma_p
## ENSG00000143333 1.736e-09   1.99000   6.492 -0.2393 0.0002045
## ENSG00000120436 8.155e-04  -0.09402   2.825 -3.3940 0.0227200
## ENSG00000099985 3.300e-10   4.85000   5.981  0.5318 0.0003530
## ENSG00000137834 1.850e-03  -0.02053   3.616 -2.8330 0.0070230
## ENSG00000177807 3.880e-03   2.09500   2.052 -3.9700 0.0748700
## ENSG00000114115 8.210e-03   0.38370   3.042 -3.1270 0.0163500
##                 limma_adjp_fdr deseq_adjp_fdr edger_adjp_fdr
## ENSG00000143333      2.282e-01      7.363e-05      4.147e-06
## ENSG00000120436      2.438e-01      1.888e-01      5.294e-02
## ENSG00000099985      2.282e-01      6.505e-06      1.314e-06
## ENSG00000137834      2.282e-01      2.751e-01      8.499e-02
## ENSG00000177807      3.169e-01      1.816e-01      1.213e-01
## ENSG00000114115      2.322e-01      3.192e-01      1.662e-01
##                 basic_adjp_fdr lfc_meta   lfc_var lfc_varbymed    p_meta
## ENSG00000143333      4.805e-01    1.507 2.720e-03    1.804e-03 6.818e-05
## ENSG00000120436      4.808e-01    1.426 2.011e-03    1.410e-03 1.003e-02
## ENSG00000099985      4.805e-01    1.273 1.195e-02    9.383e-03 1.177e-04
## ENSG00000137834      4.805e-01    1.113 2.700e-03    2.427e-03 8.914e-03
## ENSG00000177807      5.597e-01    1.008 5.935e-03    5.887e-03 2.815e-02
## ENSG00000114115      4.928e-01    1.079 1.284e-02    1.189e-02 1.774e-02
##                     p_var
## ENSG00000143333 1.394e-08
## ENSG00000120436 1.290e-04
## ENSG00000099985 4.154e-08
## ENSG00000137834 6.684e-05
## ENSG00000177807 1.638e-03
## ENSG00000114115 1.060e-04

d110_ma <- extract_de_plots(d110_pairwise, p=0.1, p_type="raw")$ma$plot
d110_ma

plot_histogram(d110_table[["data"]][["sh_vs_chr"]][, c("deseq_p")])

Try building a donor*state interaction model.

3.2 Model matrices

First make sure that self-healing is the reference.

Then make some designs of potential interest.

data <- edgeR::DGEList(exprs(hs_inf_filt))
data <- edgeR::calcNormFactors(data)
design <- pData(hs_inf_filt)

design[["condition"]] <- relevel(design[["condition"]], ref="sh")
design[["donor"]] <- as.factor(design[["donor"]])
design[["donor"]] <- relevel(design[["donor"]], ref="d107")

chsh <- model.matrix(object=~0+condition, data=design)
head(chsh) ## Looks good to me.

##               conditionsh conditionchr
## chr_5430_d108           0            1
## chr_5397_d108           0            1
## chr_2504_d108           0            1
## sh_2272_d108            1            0
## sh_1022_d108            1            0
## sh_2189_d108            1            0

donors <- model.matrix(object=~0+donor, data=design)
head(donors)

##               donord107 donord108 donord110
## chr_5430_d108         0         1         0
## chr_5397_d108         0         1         0
## chr_2504_d108         0         1         0
## sh_2272_d108          0         1         0
## sh_1022_d108          0         1         0
## sh_2189_d108          0         1         0

cond_donor_inter <- model.matrix(object=~condition+donor+condition:donor, data=design)
head(cond_donor_inter)

##               (Intercept) conditionchr donord108 donord110
## chr_5430_d108           1            1         1         0
## chr_5397_d108           1            1         1         0
## chr_2504_d108           1            1         1         0
## sh_2272_d108            1            0         1         0
## sh_1022_d108            1            0         1         0
## sh_2189_d108            1            0         1         0
##               conditionchr:donord108 conditionchr:donord110
## chr_5430_d108                      1                      0
## chr_5397_d108                      1                      0
## chr_2504_d108                      1                      0
## sh_2272_d108                       0                      0
## sh_1022_d108                       0                      0
## sh_2189_d108                       0                      0

## Why are there no conditionchr:donor107 conditionsh:donor107 columns?
## I know that if I relevel the donor factor to set the reference to another donor, that changes
## but this still does not make sense to me.

## I think the above model should be the same as:
cond_donor_interv2 <- model.matrix(object=~condition*donor, data=design)
head(cond_donor_interv2)

##               (Intercept) conditionchr donord108 donord110
## chr_5430_d108           1            1         1         0
## chr_5397_d108           1            1         1         0
## chr_2504_d108           1            1         1         0
## sh_2272_d108            1            0         1         0
## sh_1022_d108            1            0         1         0
## sh_2189_d108            1            0         1         0
##               conditionchr:donord108 conditionchr:donord110
## chr_5430_d108                      1                      0
## chr_5397_d108                      1                      0
## chr_2504_d108                      1                      0
## sh_2272_d108                       0                      0
## sh_1022_d108                       0                      0
## sh_2189_d108                       0                      0

testthat::expect_equal(cond_donor_inter, cond_donor_interv2)

## So as I understand it, if I do a test of coef='conditionchr', this is looking at
## chr vs. sh
## If I do coef='conditionchr:donor110' this is the donor effect for chr/sh for donor110?
## If this is true, then how do I get this for donor107?
## Perhaps my understanding is backwards?

donor_cond_inter <- model.matrix(object=~donor+condition+donor:condition, data=design)
donor_cond_interv2 <- model.matrix(object=~donor*condition, data=design)
testthat::expect_equal(donor_cond_inter, donor_cond_interv2)
head(donor_cond_inter)

##               (Intercept) donord108 donord110 conditionchr
## chr_5430_d108           1         1         0            1
## chr_5397_d108           1         1         0            1
## chr_2504_d108           1         1         0            1
## sh_2272_d108            1         1         0            0
## sh_1022_d108            1         1         0            0
## sh_2189_d108            1         1         0            0
##               donord108:conditionchr donord110:conditionchr
## chr_5430_d108                      1                      0
## chr_5397_d108                      1                      0
## chr_2504_d108                      1                      0
## sh_2272_d108                       0                      0
## sh_1022_d108                       0                      0
## sh_2189_d108                       0                      0

3.3 Limma implementation

voom <- limma::voom(counts=data, design=cond_donor_inter,
                    normalize.method="quant", plot=TRUE)

fitted <- limma::lmFit(object=voom, design=cond_donor_inter)
bayes <- limma::eBayes(fitted)
result <- topTable(bayes)

## Removing intercept from test coefficients

3.4 edgeR implementation

disp <- edgeR::estimateDisp(data, design=cond_donor_inter)
fit <- edgeR::glmQLFit(disp, cond_donor_inter)

## I think this is my global chr/sh
chr_sh_qlf <- edgeR::glmQLFTest(fit, coef="conditionchr")
chr_sh_result <- edgeR::topTags(chr_sh_qlf)
summary(chr_sh_result$table)

##      logFC            logCPM           F            PValue        
##  Min.   :-0.281   Min.   :1.40   Min.   :12.5   Min.   :0.000285  
##  1st Qu.: 0.395   1st Qu.:2.05   1st Qu.:13.2   1st Qu.:0.001452  
##  Median : 0.638   Median :3.48   Median :13.8   Median :0.001757  
##  Mean   : 0.581   Mean   :3.56   Mean   :14.7   Mean   :0.001629  
##  3rd Qu.: 0.899   3rd Qu.:4.78   3rd Qu.:14.5   3rd Qu.:0.002079  
##  Max.   : 1.035   Max.   :6.06   Max.   :20.9   Max.   :0.002615  
##       FDR       
##  Min.   :0.924  
##  1st Qu.:0.924  
##  Median :0.924  
##  Mean   :0.924  
##  3rd Qu.:0.924  
##  Max.   :0.924

## This should be the interaction for chr/sh and donor110.
colnames(cond_donor_inter)[5]

## [1] "conditionchr:donord108"

d110_ch_qlf <- edgeR::glmQLFTest(fit, coef="conditionchr:donord110")
d110_ch_result <- edgeR::topTags(d110_ch_qlf)
summary(d110_ch_result$table)

##      logFC            logCPM            F            PValue        
##  Min.   :-1.305   Min.   :0.762   Min.   :10.6   Min.   :0.000729  
##  1st Qu.:-0.659   1st Qu.:2.330   1st Qu.:10.9   1st Qu.:0.001731  
##  Median :-0.484   Median :3.884   Median :12.9   Median :0.002407  
##  Mean   :-0.393   Mean   :3.428   Mean   :13.0   Mean   :0.002769  
##  3rd Qu.:-0.395   3rd Qu.:4.252   3rd Qu.:13.9   3rd Qu.:0.004282  
##  Max.   : 1.133   Max.   :5.837   Max.   :17.0   Max.   :0.004813  
##       FDR   
##  Min.   :1  
##  1st Qu.:1  
##  Median :1  
##  Mean   :1  
##  3rd Qu.:1  
##  Max.   :1

3.5 DESeq2 implementation

summarized <- DESeq2::DESeqDataSetFromMatrix(countData=exprs(hs_inf_filt), colData=design,
                                             design=~condition*donor)

## converting counts to integer mode

dataset <- DESeq2::DESeqDataSet(se=summarized, design=~condition*donor)
run <- DESeq2::DESeq(dataset)

## estimating size factors

## estimating dispersions

## gene-wise dispersion estimates

## mean-dispersion relationship

## final dispersion estimates

## fitting model and testing

DESeq2::resultsNames(run)

## [1] "Intercept"              "condition_chr_vs_sh"   
## [3] "donor_d108_vs_d107"     "donor_d110_vs_d107"    
## [5] "conditionchr.donord108" "conditionchr.donord110"

res_chr_sh <- DESeq2::results(run, name="condition_chr_vs_sh")
summary(res_chr_sh)

##       Length        Class         Mode 
##            6 DESeqResults           S4

head(res_chr_sh)

## log2 fold change (MLE): condition chr vs sh 
## Wald test p-value: condition chr vs sh 
## DataFrame with 6 rows and 6 columns
##                         baseMean      log2FoldChange             lfcSE
##                        <numeric>           <numeric>         <numeric>
## ENSG00000000419 214.886619915588 -0.0934303786250222 0.123003930776459
## ENSG00000000457 318.115512380986  -0.068815027486115 0.120178565849457
## ENSG00000000460 102.031345099905   0.197578912145072 0.203833151567596
## ENSG00000000938 5676.81912456484  0.0269256146826769 0.197204584725482
## ENSG00000000971 53.5624271330181   0.344010769796457 0.293375111918846
## ENSG00000001036 460.076692835547  0.0156360632261425 0.148896600722684
##                               stat            pvalue              padj
##                          <numeric>         <numeric>         <numeric>
## ENSG00000000419 -0.759572300130944 0.447510281800168 0.997804083570585
## ENSG00000000457 -0.572606496006257 0.566911160357569 0.997804083570585
## ENSG00000000460  0.969316868358142 0.332387115533936 0.997804083570585
## ENSG00000000938  0.136536453856581 0.891397208363027 0.997804083570585
## ENSG00000000971   1.17259697847723 0.240957461669005 0.997804083570585
## ENSG00000001036  0.105012895863648 0.916365575881342 0.997804083570585

## Maybe my thinking is just a little wrong: is this the interaction of chr/sh with d110/d107?
## That would make more sense I think.
res_donor110_chrsh <- DESeq2::results(run, name="conditionchr.donord110")
summary(res_donor110_chrsh)

##       Length        Class         Mode 
##            6 DESeqResults           S4

head(res_donor110_chrsh)

## log2 fold change (MLE): conditionchr.donord110 
## Wald test p-value: conditionchr.donord110 
## DataFrame with 6 rows and 6 columns
##                         baseMean     log2FoldChange             lfcSE
##                        <numeric>          <numeric>         <numeric>
## ENSG00000000419 214.886619915588  0.295162897632011 0.172516639204654
## ENSG00000000457 318.115512380986  0.108194027119009 0.172637547612825
## ENSG00000000460 102.031345099905  0.376710409795487 0.293516835681214
## ENSG00000000938 5676.81912456484 0.0530822519196338 0.278697424287232
## ENSG00000000971 53.5624271330181 -0.388785003193934 0.436191950074647
## ENSG00000001036 460.076692835547  0.106576023570227 0.206872518783625
##                               stat             pvalue              padj
##                          <numeric>          <numeric>         <numeric>
## ENSG00000000419   1.71092422732548 0.0870951015388801 0.999936246663049
## ENSG00000000457  0.626712025368067  0.530848019781775 0.999936246663049
## ENSG00000000460    1.2834371456792  0.199338967059447 0.999936246663049
## ENSG00000000938   0.19046552746367  0.848944353796083 0.999936246663049
## ENSG00000000971 -0.891316318715647  0.372759496477823 0.999936246663049
## ENSG00000001036  0.515177289844371  0.606429137146667 0.999936246663049

res_donor108_chrsh <- DESeq2::results(run, name="conditionchr.donord108")
summary(res_donor110_chrsh)

##       Length        Class         Mode 
##            6 DESeqResults           S4

head(res_donor110_chrsh)

## log2 fold change (MLE): conditionchr.donord110 
## Wald test p-value: conditionchr.donord110 
## DataFrame with 6 rows and 6 columns
##                         baseMean     log2FoldChange             lfcSE
##                        <numeric>          <numeric>         <numeric>
## ENSG00000000419 214.886619915588  0.295162897632011 0.172516639204654
## ENSG00000000457 318.115512380986  0.108194027119009 0.172637547612825
## ENSG00000000460 102.031345099905  0.376710409795487 0.293516835681214
## ENSG00000000938 5676.81912456484 0.0530822519196338 0.278697424287232
## ENSG00000000971 53.5624271330181 -0.388785003193934 0.436191950074647
## ENSG00000001036 460.076692835547  0.106576023570227 0.206872518783625
##                               stat             pvalue              padj
##                          <numeric>          <numeric>         <numeric>
## ENSG00000000419   1.71092422732548 0.0870951015388801 0.999936246663049
## ENSG00000000457  0.626712025368067  0.530848019781775 0.999936246663049
## ENSG00000000460    1.2834371456792  0.199338967059447 0.999936246663049
## ENSG00000000938   0.19046552746367  0.848944353796083 0.999936246663049
## ENSG00000000971 -0.891316318715647  0.372759496477823 0.999936246663049
## ENSG00000001036  0.515177289844371  0.606429137146667 0.999936246663049

res_donor110_vs_107 <- DESeq2::results(run, name="donor_d110_vs_d107")
summary(res_donor110_vs_107)

##       Length        Class         Mode 
##            6 DESeqResults           S4

head(res_donor110_vs_107)

## log2 fold change (MLE): donor d110 vs d107 
## Wald test p-value: donor d110 vs d107 
## DataFrame with 6 rows and 6 columns
##                         baseMean     log2FoldChange             lfcSE
##                        <numeric>          <numeric>         <numeric>
## ENSG00000000419 214.886619915588 -0.113766010123853 0.120483955620932
## ENSG00000000457 318.115512380986 -0.456556848180769 0.121122399221616
## ENSG00000000460 102.031345099905   -0.5010401182924 0.210579406027488
## ENSG00000000938 5676.81912456484  0.505694298463598 0.197026331645787
## ENSG00000000971 53.5624271330181  -0.84242480527373  0.30875937518954
## ENSG00000001036 460.076692835547  0.568390991546206 0.145482651751353
##                               stat               pvalue
##                          <numeric>            <numeric>
## ENSG00000000419 -0.944241990873748    0.345046001807526
## ENSG00000000457  -3.76938411982257 0.000163650866259048
## ENSG00000000460   -2.3793405430491   0.0173436449623682
## ENSG00000000938   2.56663307000372   0.0102691214578703
## ENSG00000000971  -2.72841854520721  0.00636388048892617
## ENSG00000001036   3.90693312710339 9.34750101766441e-05
##                                 padj
##                            <numeric>
## ENSG00000000419    0.476062375176788
## ENSG00000000457 0.000739090391073861
## ENSG00000000460   0.0422654009174772
## ENSG00000000938   0.0271711897470485
## ENSG00000000971   0.0181060512075929
## ENSG00000001036 0.000448806683671478

3.6 No-batch data

In the following block, I will take the comparisons performed without any batch in the model/adjustment and use them to search for shared/unique genes among the self-healing vs. uninfected and the chronic vs. uninfected.

3.6.1 Review nobatch pca

hs_pairwise_nobatch_pca <- sm(plot_pca(hs_inf_filt, convert="cpm", transform="log2"))
hs_pairwise_nobatch_pca$plot

hs_pairwise_nobatch <- sm(all_pairwise(hs_uninf_filt, model_batch=FALSE,
                                       do_ebseq=FALSE))

excel_file <- glue::glue("excel/{rundate}_hs_infect_nobatch_contr-v{ver}.xlsx")
hs_combined_nobatch <- sm(combine_de_tables(
  hs_pairwise_nobatch,
  excel=excel_file,
  keepers=keepers))

excel_file <- glue::glue("excel/{rundate}_hs_infect_nobatch_sig-v{ver}.xlsx")
hs_sig_nobatch <- sm(extract_significant_genes(
  hs_combined_nobatch,
  excel=excel_file))
hs_sig_nobatch$deseq$counts

##        change_counts_up change_counts_down
## sh_nil              926                173
## ch_nil              927                167
## ch_sh                 0                  0

limma_deseq_order <- rank_order_scatter(
  hs_pairwise_nobatch,
  first_type="limma", second_type="deseq",
  first_table=1, second_table=1)
limma_deseq_order$plot

up_lst <- list(
  "sh_up" = rownames(hs_sig_nobatch[["deseq"]][["ups"]][["sh_nil"]]),
  "ch_up" = rownames(hs_sig_nobatch[["deseq"]][["ups"]][["ch_nil"]]))
nobatch_up_venn <- Vennerable::Venn(Sets=up_lst)
Vennerable::plot(nobatch_up_venn, doWeights=FALSE)

## Maria-Adelaida and Najib asked about comparing the following pair of data against
## The intersection of a bunch of stuff later...  So don't forget this!
## This gives me the genes only up in self vs. nil
nobatch_up_sh_solo <- nobatch_up_venn@IntersectionSets[["10"]]
## and ch vs nil
nobatch_up_ch_solo <- nobatch_up_venn@IntersectionSets[["01"]]

down_lst <- list(
  "sh_down" = rownames(hs_sig_nobatch[["deseq"]][["downs"]][["sh_nil"]]),
  "ch_down" = rownames(hs_sig_nobatch[["deseq"]][["downs"]][["ch_nil"]]))
nobatch_down_venn <- Vennerable::Venn(Sets=down_lst)
Vennerable::plot(nobatch_down_venn, doWeights=FALSE)

3.7 Add patient to the model

Repeat the previous set of analyses with d107/108/110 in the model.

3.7.1 Review pca

hs_pairwise_batch_pca <- sm(plot_pca(hs_inf_filt, batch="limma", convert="cpm", transform="log2"))
hs_pairwise_batch_pca$plot

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

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

hs_pairwise_batch <- sm(all_pairwise(hs_uninf_filt,
                                     model_batch=TRUE, do_ebseq=FALSE))

excel_file <- glue::glue("excel/{rundate}_hs_infect_patbatch_contr-v{ver}.xlsx")
hs_combined_batch <- sm(combine_de_tables(
  hs_pairwise_batch,
  excel=excel_file,
  keepers=keepers))

excel_file <- glue::glue("excel/{rundate}_hs_infect_patbatch_sig-v{ver}.xlsx")
hs_sig_batch <- sm(extract_significant_genes(
  hs_combined_batch,
  excel=excel_file))
hs_sig_batch[["deseq"]][["counts"]]

##        change_counts_up change_counts_down
## sh_nil              906                298
## ch_nil              932                279
## ch_sh                 0                  0

up_lst <- list(
  "sh_up" = rownames(hs_sig_batch[["deseq"]][["ups"]][["sh_nil"]]),
  "ch_up" = rownames(hs_sig_batch[["deseq"]][["ups"]][["ch_nil"]]))
batch_up_venn <- Vennerable::Venn(Sets=up_lst)
summary(batch_up_venn@IntersectionSets)

##    Length Class  Mode     
## 00   0    -none- NULL     
## 10  56    -none- character
## 01  82    -none- character
## 11 850    -none- character

Vennerable::plot(batch_up_venn, doWeights=FALSE)

down_lst <- list(
  "sh_down" = rownames(hs_sig_batch[["deseq"]][["downs"]][["sh_nil"]]),
  "ch_down" = rownames(hs_sig_batch[["deseq"]][["downs"]][["ch_nil"]]))
batch_down_venn <- Vennerable::Venn(Sets=down_lst)
Vennerable::plot(batch_down_venn, doWeights=FALSE)

summary(batch_down_venn@IntersectionSets)

##    Length Class  Mode     
## 00   0    -none- NULL     
## 10  65    -none- character
## 01  46    -none- character
## 11 233    -none- character

similar <- compare_de_results(hs_combined_nobatch, hs_combined_batch,
                              cor_method="spearman")

## Testing method: limma.

## Adding method: limma to the set.

## Testing method: deseq.

## Adding method: deseq to the set.

## Testing method: edger.

## Adding method: edger to the set.

## Testing method: ebseq.

## The first datum is missing method: ebseq.

## Testing method: basic.

## Adding method: basic to the set.

## Warning in cor(x = fs[["x"]], y = fs[["y"]], method = cor_method): the
## standard deviation is zero

## Warning in cor(x = fs[["x"]], y = fs[["y"]], method = cor_method): the
## standard deviation is zero

similar[["result"]][["deseq"]]

## $sh_nil
## $sh_nil$logfc
## [1] 0.9916
## 
## $sh_nil$p
## [1] 0.9353
## 
## $sh_nil$adjp
## [1] 0.9353
## 
## 
## $ch_nil
## $ch_nil$logfc
## [1] 0.9925
## 
## $ch_nil$p
## [1] 0.9374
## 
## $ch_nil$adjp
## [1] 0.9374
## 
## 
## $ch_sh
## $ch_sh$logfc
## [1] 0.9873
## 
## $ch_sh$p
## [1] 0.9304
## 
## $ch_sh$adjp
## [1] 0.06891

3.8 Write an excel workbook using the batch data.

The following block writes out the unique/shared genes observed among the contrasts which included donor in the model.

kept_columns <- c("ensembltranscriptid", "ensemblgeneid", "description",
                  "deseq_logfc", "deseq_adjp")
start_data <- hs_sig_batch[["deseq"]]
sh_up_solo_genes <- batch_up_venn@IntersectionSets[["10"]]
ch_up_solo_genes <- batch_up_venn@IntersectionSets[["01"]]
shch_up_shared_genes <- batch_up_venn@IntersectionSets[["11"]]
sh_down_solo_genes <- batch_down_venn@IntersectionSets[["10"]]
ch_down_solo_genes <- batch_down_venn@IntersectionSets[["01"]]
shch_down_shared_genes <- batch_down_venn@IntersectionSets[["11"]]

xls_result <- write_xls(
  data=start_data[["ups"]][["sh_nil"]][sh_up_solo_genes, kept_columns],
  sheet="sh_up_solo")
xls_result <- write_xls(
  data=start_data[["ups"]][["ch_nil"]][ch_up_solo_genes, kept_columns],
  sheet="ch_up_solo", wb=xls_result[["workbook"]])
xls_result <- write_xls(
  data=start_data[["ups"]][["ch_nil"]][shch_up_shared_genes, kept_columns],
  sheet="shch_up_shared", wb=xls_result[["workbook"]])
xls_result <- write_xls(
  data=start_data[["downs"]][["sh_nil"]][sh_down_solo_genes, kept_columns],
  sheet="sh_down_solo")
xls_result <- write_xls(
  data=start_data[["downs"]][["ch_nil"]][ch_down_solo_genes, kept_columns],
  sheet="ch_down_solo", wb=xls_result[["workbook"]])
xls_result <- write_xls(
  data=start_data[["downs"]][["ch_nil"]][shch_down_shared_genes, kept_columns],
  sheet="shch_down_shared", wb=xls_result[["workbook"]],
  excel=paste0("excel/figure_5a_stuff-v", ver, ".xlsx"))

## Saving to: excel/figure_5a_stuff-v20180822.xlsx

3.9 Add ssva into the mix

Repeat, this time attmepting to tamp down the variance by person.

3.9.1 Review ssva

hs_pairwise_ssva_pca <- sm(plot_pca(hs_inf_filt, batch="ssva", norm="quant", transform="log2"))
hs_pairwise_ssva_pca$plot

hs_pairwise_ssva <- sm(all_pairwise(hs_uninf_filt, model_batch="ssva",
                                    parallel=FALSE, do_ebseq=FALSE))

excel_file <- glue::glue("excel/{rundate}_hs_infect_fsva_contr-v{ver}.xlsx")
hs_combined_ssva <- sm(combine_de_tables(
  hs_pairwise_ssva,
  excel=excel_file,
  keepers=keepers))

excel_file <- glue::glue("excel/{rundate}_hs_infect_fsva_sig-v{ver}.xlsx")
hs_sig_ssva <- sm(extract_significant_genes(
  hs_combined_ssva,
  excel=excel_file))
hs_sig_ssva$deseq$counts

##        change_counts_up change_counts_down
## sh_nil              805                507
## ch_nil              991                553
## ch_sh                 0                  0

up_lst <- list(
  "sh_up" = rownames(hs_sig_ssva[["deseq"]][["ups"]][["sh_nil"]]),
  "ch_up" = rownames(hs_sig_ssva[["deseq"]][["ups"]][["ch_nil"]]))
ssva_up_venn <- Vennerable::Venn(Sets=up_lst)
summary(ssva_up_venn@IntersectionSets)

##    Length Class  Mode     
## 00   0    -none- NULL     
## 10  38    -none- character
## 01 224    -none- character
## 11 767    -none- character

Vennerable::plot(ssva_up_venn, doWeights=FALSE)

down_lst <- list(
  "sh_down" = rownames(hs_sig_ssva[["deseq"]][["downs"]][["sh_nil"]]),
  "ch_down" = rownames(hs_sig_ssva[["deseq"]][["downs"]][["ch_nil"]]))
ssva_down_venn <- Vennerable::Venn(Sets=down_lst)
Vennerable::plot(ssva_down_venn, doWeights=FALSE)

summary(ssva_down_venn@IntersectionSets)

##    Length Class  Mode     
## 00   0    -none- NULL     
## 10  33    -none- character
## 01  79    -none- character
## 11 474    -none- character

similar <- sm(compare_de_results(hs_combined_nobatch, hs_combined_ssva,
                                 cor_method="spearman"))
similar$result$deseq

## $sh_nil
## $sh_nil$logfc
## [1] 0.2856
## 
## $sh_nil$p
## [1] 0.008603
## 
## $sh_nil$adjp
## [1] 0.008606
## 
## 
## $ch_nil
## $ch_nil$logfc
## [1] 0.356
## 
## $ch_nil$p
## [1] 0.03392
## 
## $ch_nil$adjp
## [1] 0.03392
## 
## 
## $ch_sh
## $ch_sh$logfc
## [1] 0.9571
## 
## $ch_sh$p
## [1] 0.8855
## 
## $ch_sh$adjp
## [1] 0.06088

similar <- sm(compare_de_results(hs_combined_batch, hs_combined_ssva,
                                 cor_method="spearman"))
similar$result$deseq

## $sh_nil
## $sh_nil$logfc
## [1] 0.2982
## 
## $sh_nil$p
## [1] 0.1252
## 
## $sh_nil$adjp
## [1] 0.1252
## 
## 
## $ch_nil
## $ch_nil$logfc
## [1] 0.3653
## 
## $ch_nil$p
## [1] 0.1417
## 
## $ch_nil$adjp
## [1] 0.1417
## 
## 
## $ch_sh
## $ch_sh$logfc
## [1] 0.9735
## 
## $ch_sh$p
## [1] 0.9305
## 
## $ch_sh$adjp
## [1] 0.9305

3.10 fsva

Repeat again using fsva.

3.10.1 Review fsva

hs_pairwise_fsva_pca <- sm(plot_pca(hs_inf_filt, batch="fsva", norm="quant", transform="log2"))
hs_pairwise_fsva_pca$plot

hs_pairwise_fsva <- sm(all_pairwise(hs_uninf_filt, model_batch="fsva",
                                    parallel=FALSE, do_ebseq=FALSE))

excel_file <- glue::glue("excel/{rundate}_hs_infect_fsva_contr-v{ver}.xlsx")
hs_combined_fsva <- sm(combine_de_tables(
  hs_pairwise_fsva,
  excel=excel_file,
  keepers=keepers))

excel_file <- glue::glue("excel/{rundate}_hs_infect_fsva_sig-v{ver}.xlsx")
hs_sig_fsva <- sm(extract_significant_genes(
  hs_combined_fsva,
  excel=excel_file))

up_lst <- list(
  "sh_up" = rownames(hs_sig_fsva[["deseq"]][["ups"]][["sh_nil"]]),
  "ch_up" = rownames(hs_sig_fsva[["deseq"]][["ups"]][["ch_nil"]]))
fsva_up_venn <- Vennerable::Venn(Sets=up_lst)
summary(fsva_up_venn@IntersectionSets)

##    Length Class  Mode     
## 00   0    -none- NULL     
## 10  64    -none- character
## 01  96    -none- character
## 11 864    -none- character

Vennerable::plot(fsva_up_venn, doWeights=FALSE)

down_lst <- list(
  "sh_down" = rownames(hs_sig_fsva[["deseq"]][["downs"]][["sh_nil"]]),
  "ch_down" = rownames(hs_sig_fsva[["deseq"]][["downs"]][["ch_nil"]]))
fsva_down_venn <- Vennerable::Venn(Sets=down_lst)
Vennerable::plot(fsva_down_venn, doWeights=FALSE)

summary(fsva_down_venn@IntersectionSets)

##    Length Class  Mode     
## 00   0    -none- NULL     
## 10  70    -none- character
## 01  45    -none- character
## 11 225    -none- character

##hs_sig_fsva$deseq$counts$common_solos_fsva <- sm(subset_significants(hs_sig_fsva))
similar <- sm(compare_de_results(hs_combined_nobatch, hs_combined_fsva,
                                 cor_method="spearman"))
similar$result$deseq

## $sh_nil
## $sh_nil$logfc
## [1] 0.9937
## 
## $sh_nil$p
## [1] 0.922
## 
## $sh_nil$adjp
## [1] 0.922
## 
## 
## $ch_nil
## $ch_nil$logfc
## [1] 0.9937
## 
## $ch_nil$p
## [1] 0.9206
## 
## $ch_nil$adjp
## [1] 0.9206
## 
## 
## $ch_sh
## $ch_sh$logfc
## [1] 0.9743
## 
## $ch_sh$p
## [1] 0.9012
## 
## $ch_sh$adjp
## [1] 0.06275

similar <- sm(compare_de_results(hs_combined_ssva, hs_combined_fsva,
                                 cor_method="spearman"))
similar$result$deseq

## $sh_nil
## $sh_nil$logfc
## [1] 0.2808
## 
## $sh_nil$p
## [1] 0.1256
## 
## $sh_nil$adjp
## [1] 0.1256
## 
## 
## $ch_nil
## $ch_nil$logfc
## [1] 0.3348
## 
## $ch_nil$p
## [1] 0.1355
## 
## $ch_nil$adjp
## [1] 0.1355
## 
## 
## $ch_sh
## $ch_sh$logfc
## [1] 0.9896
## 
## $ch_sh$p
## [1] 0.9686
## 
## $ch_sh$adjp
## [1] 0.9686

3.11 Try with the combat modified data

Repeat once again, this time using combat to try to limit the contribution of the strain to the data. I do not think we will ever use this set of contrasts, so I will deactivate it but leave it here if it is required later.

old_condition <- hs_uninf$design$condition
names(old_condition) <- hs_uninf$design$sampleid
new_condition <- paste0(hs_uninf$design$state, '_', hs_uninf$design$donor)
hs_uninf_recond <- set_expt_factors(hs_uninf_filt,
                                    batch="pathogenstrain",
                                    condition=new_condition)
combat_input <- normalize_expt(hs_uninf_recond, norm="quant", batch="combat_scale")

## This function will replace the expt$expressionset slot with:

## combat_scale(quant(data))

## It backs up the current data into a slot named:
##  expt$backup_expressionset. It will also save copies of each step along the way
##  in expt$normalized with the corresponding libsizes. Keep the libsizes in mind
##  when invoking limma.  The appropriate libsize is the non-log(cpm(normalized)).
##  This is most likely kept at:
##  'new_expt$normalized$intermediate_counts$normalization$libsizes'
##  A copy of this may also be found at:
##  new_expt$best_libsize

## Filter is false, this should likely be set to something, good
##  choices include cbcb, kofa, pofa (anything but FALSE).  If you want this to
##  stay FALSE, keep in mind that if other normalizations are performed, then the
##  resulting libsizes are likely to be strange (potentially negative!)

## Leaving the data in its current base format, keep in mind that
##  some metrics are easier to see when the data is log2 transformed, but
##  EdgeR/DESeq do not accept transformed data.

## Leaving the data unconverted.  It is often advisable to cpm/rpkm
##  the data to normalize for sampling differences, keep in mind though that rpkm
##  has some annoying biases, and voom() by default does a cpm (though hpgl_voom()
##  will try to detect this).

## Step 1: not doing count filtering.

## Step 2: normalizing the data with quant.

## Step 3: not converting the data.

## Step 4: not transforming the data.

## Step 5: doing batch correction with combat_scale.

## Note to self:  If you get an error like 'x contains missing values' The data has too many 0's and needs a stronger low-count filter applied.

## batch_counts: Before batch correction, 407 entries 0<=x<1.

## batch_counts: Before batch correction, 2 entries are >= 0.

## Passing off to all_adjusters.

## Found arglist convert! raw

## The detected scale of the data is: raw.

## Doing a raw conversion with a filter with 2 threshold.

## batch_counts: Before batch correction, 407 entries 0<=x<1.

## batch_counts: Before batch correction, 2 entries are >= 0.

## The be method chose 3 surrogate variable(s).

## The most likely error in sva::empirical.controls() is a call to density in irwsva.build. Setting control_likelihoods to zero and using unsupervised sva.

## Warning in all_adjusters(count_table, design = design, estimate_type =
## batch, : It is highly likely that the underlying reason for this error is
## too many 0's in the dataset, please try doing a filtering of the data and
## retry.

## batch_counts: Using combat with a prior and with scaling.

## The number of elements which are < 0 after batch correction is: 6181

## The variable low_to_zero sets whether to change <0 values to 0 and is: FALSE

combat_input <- set_expt_conditions(combat_input, fact=old_condition)
combat_input <- set_expt_colors(combat_input, colors=c("green","blue","red"))

## The new colors are a character, changing according to condition.

3.11.1 Review combat

hs_pairwise_combat_pca <- plot_pca(combat_input)
hs_pairwise_combat_pca$plot

hs_pairwise_combatpath <- sm(all_pairwise(combat_input, model_batch=FALSE,
                                          force=TRUE))

excel_file <- glue::glue("excel/{rundate}_hs_infect_combatpath_contr-v{ver}.xlsx")
hs_combined_combatpath <- sm(combine_de_tables(
  hs_pairwise_combatpath,
  excel=excel_file,
  keepers=keepers))

excel_file <- glue::glue("excel/{rundate}_hs_infect_combatpath_sig-v{ver}.xlsx")
hs_sig_combatpath <- sm(extract_significant_genes(
  hs_combined_combatpath,
  excel=excel_file))
hs_sig_combatpath$deseq$counts

##        change_counts_up change_counts_down
## sh_nil             1518                 22
## ch_nil             2289               2056
## ch_sh               470               1091

up_lst <- list(
  "sh_up" = rownames(hs_sig_combatpath[["deseq"]][["ups"]][["sh_nil"]]),
  "ch_up" = rownames(hs_sig_combatpath[["deseq"]][["ups"]][["ch_nil"]]))
combatpath_up_venn <- Vennerable::Venn(Sets=up_lst)
summary(combatpath_up_venn@IntersectionSets)

##    Length Class  Mode     
## 00    0   -none- NULL     
## 10    0   -none- NULL     
## 01  771   -none- character
## 11 1518   -none- character

Vennerable::plot(combatpath_up_venn, doWeights=FALSE)

down_lst <- list(
  "sh_down" = rownames(hs_sig_combatpath[["deseq"]][["downs"]][["sh_nil"]]),
  "ch_down" = rownames(hs_sig_combatpath[["deseq"]][["downs"]][["ch_nil"]]))
combatpath_down_venn <- Vennerable::Venn(Sets=down_lst)
Vennerable::plot(combatpath_down_venn, doWeights=FALSE)

summary(combatpath_down_venn@IntersectionSets)

##    Length Class  Mode     
## 00    0   -none- NULL     
## 10    0   -none- NULL     
## 01 2034   -none- character
## 11   22   -none- character

similar <- sm(compare_de_results(hs_combined_nobatch, hs_combined_combatpath,
                                 cor_method="spearman"))
similar$result$deseq

## $sh_nil
## $sh_nil$logfc
## [1] -0.07967
## 
## $sh_nil$p
## [1] -0.04815
## 
## $sh_nil$adjp
## [1] -0.04815
## 
## 
## $ch_nil
## $ch_nil$logfc
## [1] -0.1114
## 
## $ch_nil$p
## [1] -0.02765
## 
## $ch_nil$adjp
## [1] -0.02765
## 
## 
## $ch_sh
## $ch_sh$logfc
## [1] 0.01295
## 
## $ch_sh$p
## [1] -0.1028
## 
## $ch_sh$adjp
## [1] 0.002389

similar <- sm(compare_de_results(hs_combined_fsva, hs_combined_combatpath,
                                 cor_method="spearman"))
similar$result$deseq

## $sh_nil
## $sh_nil$logfc
## [1] -0.08732
## 
## $sh_nil$p
## [1] -0.004519
## 
## $sh_nil$adjp
## [1] -0.004516
## 
## 
## $ch_nil
## $ch_nil$logfc
## [1] -0.1309
## 
## $ch_nil$p
## [1] 0.04731
## 
## $ch_nil$adjp
## [1] 0.04731
## 
## 
## $ch_sh
## $ch_sh$logfc
## [1] -0.04856
## 
## $ch_sh$p
## [1] -0.02353
## 
## $ch_sh$adjp
## [1] -0.0235

## OUCH!

3.12 Compare DE results

For each of the following, perform a simple DE and see what happens:

no uninfected strain as batch, try to compare each of the 3 patients chronic/self
no uninfected strain as batch, try to compare chronic/self for all
Repeat #1 above with uninfected
Repeat #2 with uninfected

3.12.1 DE: include uninfected, use strain as batch

The data used in the following is the quantile(cpm(filter())) where the condition was set to a concatenation of patient and healing state, combat was also performed, so we no longer want batch in the experimental model and also we need to pass ‘force=TRUE’ because deseq/edger will need to be coerced into accepting these modified data.

hs_inf$condition

## chr_5430_d108 chr_5397_d108 chr_2504_d108  sh_2272_d108  sh_1022_d108 
##           chr           chr           chr            sh            sh 
##  sh_2189_d108 chr_5430_d110 chr_5397_d110 chr_2504_d110  sh_2272_d110 
##            sh           chr           chr           chr            sh 
##  sh_1022_d110  sh_2189_d110 chr_5430_d107 chr_5397_d107 chr_2504_d107 
##            sh            sh           chr           chr           chr 
##  sh_2272_d107  sh_1022_d107  sh_2189_d107 
##            sh            sh            sh 
## Levels: sh chr

## Start by leaving the data relatively alone, especially noting that we do not
## have a usable batch by default.
hs_uninf_filtv2 <- hs_uninf_filt
donor_state <- paste0(hs_uninf_filtv2$design$state, "_", hs_uninf_filtv2$design$donor)
hs_uninf_filtv2 <- set_expt_factors(hs_uninf_filtv2, condition=donor_state)

uninf_patient_keepers <- list(
    "d107_chun" = c("chronic_d107", "uninfected_d107"),
    "d107_shun" = c("self_heal_d107", "uninfected_d107"),
    "d107_chsh" = c("chronic_d107", "self_heal_d107"),
    "d108_chun" = c("chronic_d108", "uninfected_d108"),
    "d108_shun" = c("self_heal_d108", "uninfected_d108"),
    "d108_chsh" = c("chronic_d108", "self_heal_d108"),
    "d110_chun" = c("chronic_d110", "uninfected_d110"),
    "d110_shun" = c("self_heal_d110", "uninfected_d110"),
    "d110_chsh" = c("chronic_d110", "self_heal_d110"))

hs_uninf_filtv2_pairwise <- sm(all_pairwise(hs_uninf_filtv2, parallel=FALSE,
                                            model_batch=FALSE, do_ebseq=FALSE))

hs_uninf_filtv2_combined <- sm(combine_de_tables(
  hs_uninf_filtv2_pairwise,
  keepers=uninf_patient_keepers,
  excel=paste0("excel/hs_infect_patient_nobatch-v", ver, ".xlsx")))

hs_uninf_filtv2_sig <- sm(extract_significant_genes(
  hs_uninf_filtv2_combined, according_to="deseq",
  excel=paste0("excel/hs_infect_patient_nobatch_sig-v", ver, ".xlsx")))

##hs_uninf_filtv3_pairwise <- all_pairwise(hs_uninf_filtv2, model_batch="svaseq", surrogates=1)
##hs_uninf_filtv3_combined <- sm(combine_de_tables(hs_uninf_filtv3_pairwise,
##                                                 keepers=uninf_patient_keepers,
##                                                 excel=paste0("excel/hs_infect_patient_fsva-v", ver, ".xlsx")))
##hs_uninf_filtv3_sig <- sm(extract_significant_genes(hs_uninf_filtv3_combined,
##                                                    excel=paste0("excel/hs_infect_patient_fsva_sig-v", ver, ".xlsx")))

3.13 Make some Venns

Now we want to look at intersections from the perspective of contrasts performed comparing the self-healing/chronic vs. uninfected for the three donors separately.

3.13.1 Perform venns of self-healing vs. uninfected

This time for each of the three donors: self-healing up vs. uninfected.

up_sig <- hs_uninf_filtv2_sig$deseq$ups

## Do this in a way which is not stupid...
comp_lst <- list(
  "donor_107" = rownames(up_sig[["d107_shun"]]),
  "donor_108" = rownames(up_sig[["d108_shun"]]),
  "donor_110" = rownames(up_sig[["d110_shun"]]))
comp_shun_up <- Vennerable::Venn(Sets=comp_lst)
up_res <- Vennerable::plot(comp_shun_up, doWeights=FALSE)

shared_shun_up <- comp_shun_up@IntersectionSets[["111"]]
de_table_shared_up_sh_first <- hs_uninf_filtv2_combined[["data"]][["d107_shun"]][shared_shun_up, ]
de_table_shared_up_sh_second <- hs_uninf_filtv2_combined[["data"]][["d108_shun"]][shared_shun_up, ]
de_table_shared_up_sh_third <- hs_uninf_filtv2_combined[["data"]][["d110_shun"]][shared_shun_up, ]
de_table_shared_up_sh_all <- merge(
  de_table_shared_up_sh_first[, c("description", "deseq_logfc", "deseq_adjp")],
  de_table_shared_up_sh_second[, c("deseq_logfc", "deseq_adjp")],
  by="row.names")
de_table_shared_up_sh_all <- merge(
  de_table_shared_up_sh_all,
  de_table_shared_up_sh_third[, c("deseq_logfc", "deseq_adjp")],
  by.x="Row.names", by.y="row.names")
rownames(de_table_shared_up_sh_all) <- de_table_shared_up_sh_all[["Row.names"]]
de_table_shared_up_sh_all <- de_table_shared_up_sh_all[, -1]
colnames(de_table_shared_up_sh_all) <- c("description", "logfc_107", "adjp_107",
                                         "logfc_108", "adjp_108", "logfc_110", "adjp_110")
write.csv(de_table_shared_up_sh_all, file="images/de_table_shared_up_sh_all.csv")

3.13.2 Perform venns of chronic vs. uninfected

This time for each of the three donors: chronic up vs. uninfected.

up_sig <- hs_uninf_filtv2_sig$deseq$ups
comp_lst <- list(
  "donor_107" = rownames(up_sig[["d107_chun"]]),
  "donor_108" = rownames(up_sig[["d108_chun"]]),
  "donor_110" = rownames(up_sig[["d110_chun"]]))
comp_chun_up <- Vennerable::Venn(Sets=comp_lst)
up_res <- Vennerable::plot(comp_chun_up, doWeights=FALSE)

shared_chun_up <- comp_chun_up@IntersectionSets[["111"]]
de_table_shared_up_ch_first <- hs_uninf_filtv2_combined[["data"]][["d107_chun"]][shared_chun_up, ]
de_table_shared_up_ch_second <- hs_uninf_filtv2_combined[["data"]][["d108_chun"]][shared_chun_up, ]
de_table_shared_up_ch_third <- hs_uninf_filtv2_combined[["data"]][["d110_chun"]][shared_chun_up, ]
de_table_shared_up_ch_all <- merge(
  de_table_shared_up_ch_first[, c("description", "deseq_logfc", "deseq_adjp")],
  de_table_shared_up_ch_second[, c("deseq_logfc", "deseq_adjp")],
  by="row.names")
de_table_shared_up_ch_all <- merge(
  de_table_shared_up_ch_all,
  de_table_shared_up_ch_third[, c("deseq_logfc", "deseq_adjp")],
  by.x="Row.names", by.y="row.names")
rownames(de_table_shared_up_ch_all) <- de_table_shared_up_ch_all[["Row.names"]]
de_table_shared_up_ch_all <- de_table_shared_up_ch_all[, -1]
colnames(de_table_shared_up_ch_all) <- c("description", "logfc_107", "adjp_107",
                                         "logfc_108", "adjp_108", "logfc_110", "adjp_110")
write.csv(de_table_shared_up_ch_all, file="images/de_table_shared_up_ch_all.csv")

3.13.3 Perform venns of self-healing vs. uninfected

This time for each of the three donors: self-healing down vs. uninfected.

down_sig <- hs_uninf_filtv2_sig$deseq$downs
comp_lst <- list(
  "donor_107" = rownames(down_sig[["d107_shun"]]),
  "donor_108" = rownames(down_sig[["d108_shun"]]),
  "donor_110" = rownames(down_sig[["d110_shun"]]))
comp_shun_down <- Vennerable::Venn(Sets=comp_lst)
down_res <- Vennerable::plot(comp_shun_down, doWeights=FALSE)

shared_shun_down <- comp_shun_down@IntersectionSets[["111"]]
de_table_shared_down_sh_first <- hs_uninf_filtv2_combined[["data"]][["d107_shun"]][shared_shun_down, ]
de_table_shared_down_sh_second <- hs_uninf_filtv2_combined[["data"]][["d108_shun"]][shared_shun_down, ]
de_table_shared_down_sh_third <- hs_uninf_filtv2_combined[["data"]][["d110_shun"]][shared_shun_down, ]
de_table_shared_down_sh_all <- merge(
  de_table_shared_down_sh_first[, c("description", "deseq_logfc", "deseq_adjp")],
  de_table_shared_down_sh_second[, c("deseq_logfc", "deseq_adjp")],
  by="row.names")
de_table_shared_down_sh_all <- merge(
  de_table_shared_down_sh_all,
  de_table_shared_down_sh_third[, c("deseq_logfc", "deseq_adjp")],
  by.x="Row.names", by.y="row.names")
rownames(de_table_shared_down_sh_all) <- de_table_shared_down_sh_all[["Row.names"]]
de_table_shared_down_sh_all <- de_table_shared_down_sh_all[, -1]
colnames(de_table_shared_down_sh_all) <- c("description", "logfc_107",
                                           "adjp_107", "logfc_108", "adjp_108",
                                           "logfc_110", "adjp_110")
write.csv(de_table_shared_down_sh_all, file="images/de_table_shared_down_sh_all.csv")

3.13.4 Perform venns of self-healing vs. uninfected

This time for each of the three donors: chronic down vs. uninfected.

down_sig <- hs_uninf_filtv2_sig$deseq$downs
comp_lst <- list(
  "donor_107" = rownames(down_sig[["d107_chun"]]),
  "donor_108" = rownames(down_sig[["d108_chun"]]),
  "donor_110" = rownames(down_sig[["d110_chun"]]))
comp_chun_down <- Vennerable::Venn(Sets=comp_lst)
down_res <- Vennerable::plot(comp_chun_down, doWeights=FALSE)

shared_chun_down <- comp_chun_down@IntersectionSets[["111"]]
de_table_shared_down_ch_first <- hs_uninf_filtv2_combined[["data"]][["d107_chun"]][shared_chun_down, ]
de_table_shared_down_ch_second <- hs_uninf_filtv2_combined[["data"]][["d108_chun"]][shared_chun_down, ]
de_table_shared_down_ch_third <- hs_uninf_filtv2_combined[["data"]][["d110_chun"]][shared_chun_down, ]
de_table_shared_down_ch_all <- merge(
  de_table_shared_down_ch_first[, c("description", "deseq_logfc", "deseq_adjp")],
  de_table_shared_down_ch_second[, c("deseq_logfc", "deseq_adjp")],
  by="row.names")
de_table_shared_down_ch_all <- merge(
  de_table_shared_down_ch_all,
  de_table_shared_down_ch_third[, c("deseq_logfc", "deseq_adjp")],
  by.x="Row.names", by.y="row.names")
rownames(de_table_shared_down_ch_all) <- de_table_shared_down_ch_all[["Row.names"]]
de_table_shared_down_ch_all <- de_table_shared_down_ch_all[, -1]
colnames(de_table_shared_down_ch_all) <- c("description", "logfc_107",
                                           "adjp_107", "logfc_108", "adjp_108",
                                           "logfc_110", "adjp_110")
write.csv(de_table_shared_down_ch_all, file="images/de_table_shared_down_ch_all.csv")

At this point, we should have a set of genes which are up/down in the self/uninfected and chronic/uninfected, kept in variables with names like: ‘shared_shun_down’ and ‘shared_chun_down’

3.14 Get Meta! Intersect the above comparisons.

Now we want a sense of what genes are shared/unique among the self-healing vs. uninfected and the chronic vs. uninfected comparisons performed above. One would assume that the most interesting genes in these sets will prove to be the the ones which are not shared.

shch_up_lst <- list(
  "up_sh" = shared_shun_up,
  "up_ch" = shared_chun_up)
shared_up <- Vennerable::Venn(Sets=shch_up_lst)
Vennerable::plot(shared_up, doWeights=FALSE)

shch_down_lst <- list(
  "up_sh" = shared_shun_down,
  "up_ch" = shared_chun_down)
shared_down <- Vennerable::Venn(Sets=shch_down_lst)
Vennerable::plot(shared_down, doWeights=FALSE)

upup_genes <- shared_up@IntersectionSets[["11"]]
upsh_notch <- shared_up@IntersectionSets[["10"]]
upch_notsh <- shared_up@IntersectionSets[["01"]]
downdown_genes <- shared_down@IntersectionSets[["11"]]
downsh_notch <- shared_down@IntersectionSets[["10"]]
downch_notsh <- shared_down@IntersectionSets[["01"]]

kept_columns <- c("ensembltranscriptid", "ensemblgeneid", "description",
                  "deseq_logfc", "deseq_adjp")
## Get the shared things in up_sh and up_ch, ergo can use either sheet
xls_result <- write_xls(data=up_sig[["d107_shun"]][upup_genes, kept_columns],
                        sheet="upsh_upch")
## Get the only-up_sh
xls_result <- write_xls(data=up_sig[["d107_shun"]][upsh_notch, kept_columns],
                        sheet="upsh_noch",
                        wb=xls_result[["workbook"]])
## Get the only-up_ch
xls_result <- write_xls(data=up_sig[["d107_chun"]][upch_notsh, kept_columns],
                        sheet="upch_nosh",
                        wb=xls_result[["workbook"]])
## Now repeat for the down: down-shared
## Get shared down down_sh down_ch
xls_result <- write_xls(data=down_sig[["d107_shun"]][downdown_genes, kept_columns],
                        sheet="downsh_downch",
                        wb=xls_result[["workbook"]])
## The down_sh only
xls_result <- write_xls(data=down_sig[["d107_shun"]][downsh_notch, kept_columns],
                        sheet="downsh_noch",
                        wb=xls_result[["workbook"]])
## The down_ch
xls_result <- write_xls(data=down_sig[["d107_chun"]][downch_notsh, kept_columns],
                        sheet="downch_nosh",
                        wb=xls_result[["workbook"]],
                        excel=paste0("excel/figure_5c_stuff-v", ver, ".xlsx"))

## Saving to: excel/figure_5c_stuff-v20180822.xlsx

3.14.1 Genes shared among the donors.

One further query: what genes are shared among the contrasts of self-healing/chronic vs. uninfected for the three donors? When doing this, we have once again to consider whether to use the nobatch/batch-in-model/sva/etc methods against our donors… Since they all agree pretty well until combat, I will arbitrarily choose fsva.

One idea suggested by Maria Adelaida was to compare the set of genes shared among d107/d108/d110 in this last comparison (which was each of the three donors separately) against the set of genes in the all-data analysis above.

3.14.1.1 Attempt to answer by the unique individual sets

In this block, I will attempt to answer the above query by intersecting the sets of genes shared among the individuals but unique to sh/chr vs. uninfected against the set of genes observed in the original sva-mediated DE analysis.

indiv_shup_unique <- shared_up@IntersectionSets[["10"]]
avg_vs_ind_shun_up <- list(
  "fsva_avg" = rownames(hs_sig_fsva[["deseq"]][["ups"]][["sh_nil"]]),
  "indiv" = indiv_shup_unique)
avg_ind_shup_venn <- Vennerable::Venn(Sets=avg_vs_ind_shun_up)
Vennerable::plot(avg_ind_shup_venn, doWeights=FALSE)

indiv_chup_unique <- shared_up@IntersectionSets[["01"]]
avg_vs_ind_chun_up <- list(
  "fsva_avg" = rownames(hs_sig_fsva[["deseq"]][["ups"]][["ch_nil"]]),
  "indiv" = indiv_shup_unique)
avg_ind_chup_venn <- Vennerable::Venn(Sets=avg_vs_ind_chun_up)
Vennerable::plot(avg_ind_chup_venn, doWeights=FALSE)

indiv_shdown_unique <- shared_down@IntersectionSets[["10"]]
avg_vs_ind_shun_down <- list(
  "fsva_avg" = rownames(hs_sig_fsva[["deseq"]][["downs"]][["sh_nil"]]),
  "indiv" = indiv_shdown_unique)
avg_ind_shdown_venn <- Vennerable::Venn(Sets=avg_vs_ind_shun_down)
Vennerable::plot(avg_ind_shdown_venn, doWeights=FALSE)

indiv_chdown_unique <- shared_down@IntersectionSets[["01"]]
avg_vs_ind_chun_down <- list(
  "fsva_avg" = rownames(hs_sig_fsva[["deseq"]][["downs"]][["ch_nil"]]),
  "indiv" = indiv_shdown_unique)
avg_ind_chdown_venn <- Vennerable::Venn(Sets=avg_vs_ind_chun_down)
Vennerable::plot(avg_ind_chdown_venn, doWeights=FALSE)

3.14.1.2 Answer by looking at the shared genes in both sets

shun_up_avg_genes <- rownames(hs_sig_batch[["deseq"]][["ups"]][["sh_nil"]])
chun_up_avg_genes <- rownames(hs_sig_batch[["deseq"]][["ups"]][["ch_nil"]])
shun_down_avg_genes <- rownames(hs_sig_batch[["deseq"]][["downs"]][["sh_nil"]])
chun_down_avg_genes <- rownames(hs_sig_batch[["deseq"]][["downs"]][["ch_nil"]])

shun_up_individual_genes <- comp_shun_up@IntersectionSets[["111"]]
chun_up_individual_genes <- comp_chun_up@IntersectionSets[["111"]]
shun_down_individual_genes <- comp_shun_down@IntersectionSets[["111"]]
chun_down_individual_genes <- comp_chun_down@IntersectionSets[["111"]]

last_shun_up_lst <- list(
  "avg_genes" = shun_up_avg_genes,
  "ind_genes" = shun_up_individual_genes)
last_shun_up_venn <- Vennerable::Venn(Sets=last_shun_up_lst)
Vennerable::plot(last_shun_up_venn, doWeights=FALSE)

last_chun_up_lst <- list(
  "avg_genes" = chun_up_avg_genes,
  "ind_genes" = chun_up_individual_genes)
last_chun_up_venn <- Vennerable::Venn(Sets=last_chun_up_lst)
Vennerable::plot(last_chun_up_venn, doWeights=FALSE)

last_shun_down_lst <- list(
  "avg_genes" = shun_down_avg_genes,
  "ind_genes" = shun_down_individual_genes)
last_shun_down_venn <- Vennerable::Venn(Sets=last_shun_down_lst)
Vennerable::plot(last_shun_down_venn, doWeights=FALSE)

last_chun_down_lst <- list(
  "avg_genes" = chun_down_avg_genes,
  "ind_genes" = chun_down_individual_genes)
last_chun_down_venn <- Vennerable::Venn(Sets=last_chun_down_lst)
Vennerable::plot(last_chun_down_venn, doWeights=FALSE)

3.14.2 DE: include uninfected, repeat with condition simplified to chronic/self

Do I think the exact same thing as in the previous comparison, but now simplify the ‘condition’ factor to just self-healing vs. chronic and see what happens.

uninf_strainbatch_qcf <- set_expt_batches(expt=hs_cds_uninf, fact="pathogenstrain")
uninf_strainbatch_qcf <- set_expt_conditions(expt=uninf_strainbatch_qcf, fact="state")
withuninf_strainbatch_pairs_chsh <- all_pairwise(
  parallel=FALSE, uninf_strainbatch_qcf,
  model_batch=FALSE, force=TRUE, do_ebseq=FALSE)
chsh_keepers <- list(
    "ch_sh" = c("chronic", "self_heal"),
    "ch_nil" = c("chronic", "uninfected"),
    "sh_nil" = c("self_heal", "uninfected"))
withuninf_strainbatch_tables_chsh <- sm(combine_de_tables(
  withuninf_strainbatch_pairs_chsh,
  keepers=chsh_keepers,
  excel=paste0("excel/withuninf_strainbatch_chsh_pairwise-v", ver, ".xlsx")))
withuninf_strainbatch_sig_chsh <- extract_significant_genes(
  withuninf_strainbatch_tables_chsh,
  excel=paste0("excel/withuninf_strainbatch_chsh_sig-v", ver, ".xlsx"))
##withuninf_strainbatch_tables_chsh$limma_plots
##withuninf_strainbatch_tables_chsh$edger_plots
##withuninf_strainbatch_tables_chsh$deseq_plots

4 Figure 5

Generate DE lists of each donor for all contrasts for PBMCs.

Venn sh/uninf vs chr/uninf 2 venn diagram up. (donor in model)
Venn sh/uninf vs chr/uninf 2 venn diagram down.
Venn Sh/uninf up genes 3 venn diagram.
Venn Sh/uninf down genes 3 venn.
Venn Chr/uninf up genes 3 venn.
Venn Chr/uninf down genes 3 venn.
2 way venn of (common up in 3 venn sh/uninf) vs. (common up in 3 venn chr/uninf)
2 way venn of (common down in 3 venn sh/uninf) vs. (common down in 3 venn chr/uninf)

I renamed these plots and am now hopelessly confused as to which is which. I think I will not run this for now but instead generate the worksheet without them and then return to this in the hopes that I can do a better job.

pp(file="images/fig_05a-sh_uninf_vs_chr_uninf_up.pdf")
Vennerable::plot(common_solos_batch$up_venn, doWeights=FALSE)
dev.off()
pp(file="images/fig_05b-sh_uninf_vs_chr_uninf_down.pdf")
Vennerable::plot(common_solos_batch$down_venn, doWeights=FALSE)
dev.off()
pp(file="images/fig_05c-sh_uninf_donors_up.pdf")
Vennerable::plot(sh_up_venn, doWeights=FALSE)
dev.off()
pp(file="images/fig_05d-sh_uninf_donors_down.pdf")
Vennerable::plot(sh_down_venn, doWeights=FALSE)
dev.off()
pp(file="images/fig_05e-chr_uninf_donors_up.pdf")
Vennerable::plot(chr_up_venn, doWeights=FALSE)
dev.off()
pp(file="images/fig_05f-chr_uninf_donors_down.pdf")
Vennerable::plot(chr_down_venn, doWeights=FALSE)
dev.off()
pp(file="images/fig_05g-up_common.pdf")
Vennerable::plot(shared_up, doWeights=FALSE)
dev.off()
pp(file="images/fig_05h-down_common.pdf")
Vennerable::plot(shared_down, doWeights=FALSE)
dev.off()

5 Try again on the parasite data

5.1 Remember our data set

lp_inf_filt <- sm(normalize_expt(lp_inf, filter=TRUE))

lp_pairwise_nobatch <- sm(all_pairwise(lp_inf_filt, model_batch=FALSE,
                                       parallel=FALSE, do_ebseq=FALSE))

excel_file <- glue::glue("excel/{rundate}_lp_infect_nobatch_contr-v{ver}.xlsx")
lp_combined_nobatch <- sm(combine_de_tables(lp_pairwise_nobatch, excel=excel_file))

excel_file <- glue::glue("excel/{rundate}_lp_infect_nobatch_sig-v{ver}.xlsx")
lp_sig_nobatch <- sm(extract_significant_genes(lp_combined_nobatch, excel=excel_file))

lp_pairwise_batch <- sm(all_pairwise(lp_inf_filt, model_batch=TRUE))
excel_file <- glue::glue("excel/{rundate}_lp_infect_batch_contr-v{ver}.xlsx")
lp_combined_batch <- sm(combine_de_tables(lp_pairwise_batch, excel=excel_file))

excel_file <- glue::glue("excel/{rundate}_lp_infect_batch_sig-v{ver}.xlsx")
lp_sig_batch <- sm(extract_significant_genes(lp_combined_batch, excel=excel_file))

lp_pairwise_ssva <- sm(all_pairwise(lp_inf_filt, model_batch="ssva"))
excel_file <- glue::glue("excel/{rundate}_lp_infect_ssva_contr-v{ver}.xlsx")
lp_combined_ssva <- sm(combine_de_tables(lp_pairwise_ssva, excel=excel_file))

excel_file <- glue::glue("excel/{rundate}_lp_infect_ssva_sig-v{ver}.xlsx")
lp_sig_ssva <- sm(extract_significant_genes(lp_combined_ssva, excel=excel_file))

lp_pairwise_fsva <- sm(all_pairwise(lp_inf_filt, model_batch="fsva"))
excel_file <- glue::glue("excel/{rundate}_lp_infect_fsva_contr-v{ver}.xlsx")
lp_combined_fsva <- sm(combine_de_tables(lp_pairwise_fsva, excel=excel_file))

excel_file <- glue::glue("excel/{rundate}_lp_infect_fsva_sig-v{ver}.xlsx")
lp_sig_fsva <- sm(extract_significant_genes(lp_combined_fsva, excel=excel_file))

pander::pander(sessionInfo())

R version 3.5.1 (2018-07-02)

Platform: x86_64-pc-linux-gnu (64-bit)

locale: LC_CTYPE=en_US.utf8, LC_NUMERIC=C, LC_TIME=en_US.utf8, LC_COLLATE=en_US.utf8, LC_MONETARY=en_US.utf8, LC_MESSAGES=en_US.utf8, LC_PAPER=en_US.utf8, LC_NAME=C, LC_ADDRESS=C, LC_TELEPHONE=C, LC_MEASUREMENT=en_US.utf8 and LC_IDENTIFICATION=C

attached base packages: parallel, stats, graphics, grDevices, utils, datasets, methods and base

other attached packages: edgeR(v.3.24.0), ruv(v.0.9.7), bindrcpp(v.0.2.2), variancePartition(v.1.12.0), scales(v.1.0.0), foreach(v.1.4.4), limma(v.3.38.3), ggplot2(v.3.1.0), hpgltools(v.2018.11), Biobase(v.2.42.0) and BiocGenerics(v.0.28.0)

loaded via a namespace (and not attached): tidyselect(v.0.2.5), lme4(v.1.1-19), RSQLite(v.2.1.1), AnnotationDbi(v.1.44.0), htmlwidgets(v.1.3), grid(v.3.5.1), BiocParallel(v.1.16.2), devtools(v.2.0.1), munsell(v.0.5.0), codetools(v.0.2-15), preprocessCore(v.1.44.0), units(v.0.6-1), withr(v.2.1.2), colorspace(v.1.3-2), GOSemSim(v.2.8.0), OrganismDbi(v.1.24.0), knitr(v.1.20), rstudioapi(v.0.8), stats4(v.3.5.1), Vennerable(v.3.1.0.9000), robustbase(v.0.93-3), DOSE(v.3.8.0), labeling(v.0.3), urltools(v.1.7.1), GenomeInfoDbData(v.1.2.0), bit64(v.0.9-7), farver(v.1.1.0), rprojroot(v.1.3-2), R6(v.2.3.0), doParallel(v.1.0.14), GenomeInfoDb(v.1.18.1), locfit(v.1.5-9.1), bitops(v.1.0-6), fgsea(v.1.8.0), gridGraphics(v.0.3-0), DelayedArray(v.0.8.0), assertthat(v.0.2.0), ggraph(v.1.0.2), nnet(v.7.3-12), enrichplot(v.1.2.0), gtable(v.0.2.0), sva(v.3.30.0), processx(v.3.2.0), rlang(v.0.3.0.1), genefilter(v.1.64.0), splines(v.3.5.1), rtracklayer(v.1.42.1), lazyeval(v.0.2.1), acepack(v.1.4.1), checkmate(v.1.8.5), europepmc(v.0.3), BiocManager(v.1.30.4), yaml(v.2.2.0), reshape2(v.1.4.3), GenomicFeatures(v.1.34.1), backports(v.1.1.2), qvalue(v.2.14.0), Hmisc(v.4.1-1), clusterProfiler(v.3.10.0), RBGL(v.1.58.1), tools(v.3.5.1), usethis(v.1.4.0), ggplotify(v.0.0.3), gplots(v.3.0.1), RColorBrewer(v.1.1-2), sessioninfo(v.1.1.1), ggridges(v.0.5.1), Rcpp(v.1.0.0), plyr(v.1.8.4), base64enc(v.0.1-3), progress(v.1.2.0), zlibbioc(v.1.28.0), purrr(v.0.2.5), RCurl(v.1.95-4.11), ps(v.1.2.1), prettyunits(v.1.0.2), rpart(v.4.1-13), viridis(v.0.5.1), cowplot(v.0.9.3), S4Vectors(v.0.20.1), SummarizedExperiment(v.1.12.0), ggrepel(v.0.8.0), cluster(v.2.0.7-1), colorRamps(v.2.3), fs(v.1.2.6), magrittr(v.1.5), data.table(v.1.11.8), openxlsx(v.4.1.0), DO.db(v.2.9), triebeard(v.0.3.0), packrat(v.0.5.0), matrixStats(v.0.54.0), pkgload(v.1.0.2), hms(v.0.4.2), evaluate(v.0.12), xtable(v.1.8-3), pbkrtest(v.0.4-7), XML(v.3.98-1.16), IRanges(v.2.16.0), gridExtra(v.2.3), testthat(v.2.0.1), compiler(v.3.5.1), biomaRt(v.2.38.0), tibble(v.1.4.2), KernSmooth(v.2.23-15), crayon(v.1.3.4), minqa(v.1.2.4), htmltools(v.0.3.6), mgcv(v.1.8-26), corpcor(v.1.6.9), snow(v.0.4-3), Formula(v.1.2-3), geneplotter(v.1.60.0), tidyr(v.0.8.2), DBI(v.1.0.0), tweenr(v.1.0.0), MASS(v.7.3-51.1), Matrix(v.1.2-15), cli(v.1.0.1), quadprog(v.1.5-5), gdata(v.2.18.0), bindr(v.0.1.1), igraph(v.1.2.2), GenomicRanges(v.1.34.0), pkgconfig(v.2.0.2), rvcheck(v.0.1.1), GenomicAlignments(v.1.18.0), foreign(v.0.8-71), plotly(v.4.8.0), xml2(v.1.2.0), annotate(v.1.60.0), XVector(v.0.22.0), stringr(v.1.3.1), callr(v.3.0.0), digest(v.0.6.18), graph(v.1.60.0), Biostrings(v.2.50.1), rmarkdown(v.1.10), fastmatch(v.1.1-0), htmlTable(v.1.12), directlabels(v.2018.05.22), Rsamtools(v.1.34.0), gtools(v.3.8.1), nloptr(v.1.2.1), nlme(v.3.1-137), jsonlite(v.1.5), desc(v.1.2.0), viridisLite(v.0.3.0), pillar(v.1.3.0), lattice(v.0.20-38), DEoptimR(v.1.0-8), httr(v.1.3.1), pkgbuild(v.1.0.2), survival(v.2.43-3), GO.db(v.3.7.0), glue(v.1.3.0), remotes(v.2.0.2), zip(v.1.0.0), UpSetR(v.1.3.3), iterators(v.1.0.10), pander(v.0.6.3), bit(v.1.1-14), ggforce(v.0.1.3), stringi(v.1.2.4), blob(v.1.1.1), DESeq2(v.1.22.1), doSNOW(v.1.0.16), latticeExtra(v.0.6-28), caTools(v.1.17.1.1), memoise(v.1.1.0) and dplyr(v.0.7.8)

message(paste0("This is hpgltools commit: ", get_git_commit()))

## If you wish to reproduce this exact build of hpgltools, invoke the following:

## > git clone http://github.com/abelew/hpgltools.git

## > git reset 242609d9ad73083ea13099760f8f0678a4befbf8

## This is hpgltools commit: Wed Dec 5 15:02:17 2018 -0500: 242609d9ad73083ea13099760f8f0678a4befbf8

this_save <- paste0(gsub(pattern="\\.Rmd", replace="", x=rmd_file), "-v", ver, ".rda.xz")
message(paste0("Saving to ", this_save))

## Saving to 03_expression_infection_20180822-v20180822.rda.xz

tmp <- sm(saveme(filename=this_save))

6 State saved to `this_save`

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

L. panamensis 20180822: Differential Expression of infected PBMCs.

atb abelew@gmail.com

2018-12-06

1 PBMC Infection Differential Expression, Infection: 20180822 Rundate: 20181206

1.1 Create data sets to compare differential expression analyses

2 The negative control

2.1 Filter the data

3 Queries from 20181205

3.1 Each individual, chr vs sh.

3.2 Model matrices

3.3 Limma implementation

3.4 edgeR implementation

3.5 DESeq2 implementation

3.6 No-batch data

3.6.1 Review nobatch pca

3.7 Add patient to the model

3.7.1 Review pca

3.8 Write an excel workbook using the batch data.

3.9 Add ssva into the mix

3.9.1 Review ssva

3.10 fsva

3.10.1 Review fsva

3.11 Try with the combat modified data

3.11.1 Review combat

3.12 Compare DE results

3.12.1 DE: include uninfected, use strain as batch

3.13 Make some Venns

3.13.1 Perform venns of self-healing vs. uninfected

3.13.2 Perform venns of chronic vs. uninfected

3.13.3 Perform venns of self-healing vs. uninfected

3.13.4 Perform venns of self-healing vs. uninfected

3.14 Get Meta! Intersect the above comparisons.

3.14.1 Genes shared among the donors.

3.14.1.1 Attempt to answer by the unique individual sets

3.14.1.2 Answer by looking at the shared genes in both sets

3.14.2 DE: include uninfected, repeat with condition simplified to chronic/self

4 Figure 5

5 Try again on the parasite data

5.1 Remember our data set

6 State saved to `this_save`

L. panamensis 20180822: Differential Expression of infected PBMCs.

atb abelew@gmail.com

2018-12-06

1 PBMC Infection Differential Expression, Infection: 20180822 Rundate: 20181206

1.1 Create data sets to compare differential expression analyses

2 The negative control

2.1 Filter the data

3 Queries from 20181205

3.1 Each individual, chr vs sh.

3.2 Model matrices

3.3 Limma implementation

3.4 edgeR implementation

3.5 DESeq2 implementation

3.6 No-batch data

3.6.1 Review nobatch pca

3.7 Add patient to the model

3.7.1 Review pca

3.8 Write an excel workbook using the batch data.

3.9 Add ssva into the mix

3.9.1 Review ssva

3.10 fsva

3.10.1 Review fsva

3.11 Try with the combat modified data

3.11.1 Review combat

3.12 Compare DE results

3.12.1 DE: include uninfected, use strain as batch

3.13 Make some Venns

3.13.1 Perform venns of self-healing vs. uninfected

3.13.2 Perform venns of chronic vs. uninfected

3.13.3 Perform venns of self-healing vs. uninfected

3.13.4 Perform venns of self-healing vs. uninfected

3.14 Get Meta! Intersect the above comparisons.

3.14.1 Genes shared among the donors.

3.14.1.1 Attempt to answer by the unique individual sets

3.14.1.2 Answer by looking at the shared genes in both sets

3.14.2 DE: include uninfected, repeat with condition simplified to chronic/self

4 Figure 5

5 Try again on the parasite data

5.1 Remember our data set

6 State saved to this_save

6 State saved to `this_save`