Chapter 9 Biopsy Quality Control

Blood, fat, muscle markers and scored by \(\log_{10}(TPM + 1)\).

9.1 Blood, fat and muscle cell markers

markers <- tibble(cell_type=c(rep("blood", 3), rep("fat", 3), rep("muscle", 3)),
                  gene_name=c("HBA1", "HBA2", "HBB",
                              "FASN", "LEP", "SCD",
                              "ACTA1", "TNNT3", "MYH1"))  %>%
  dplyr::mutate(gencode_id = get_ensembl(gene_name, rse=sanitized.dds)) %>%
  dplyr::mutate(cell_type = factor(cell_type))
## HBA1 HBA2 HBB FASN LEP SCD ACTA1 TNNT3 MYH1
knitr::kable(markers)
cell_type gene_name gencode_id
blood HBA1 ENSG00000206172.8
blood HBA2 ENSG00000188536.12
blood HBB ENSG00000244734.3
fat FASN ENSG00000169710.7
fat LEP ENSG00000174697.4
fat SCD ENSG00000099194.5
muscle ACTA1 ENSG00000143632.14
muscle TNNT3 ENSG00000130595.16
muscle MYH1 ENSG00000109061.9

9.2 Define score

For each cell type, the marker score for each sample is \(\frac{1}{n}\Sigma_{i=1}^{n}\log_{10}(TPM_i + 1)\), where \(i\) denotes genes and \(n\) is the number of genes of the cell type.

lv <- levels(markers$cell_type)
celltype_tpm <- sapply(lv, function(type) {
   id <- markers %>% dplyr::filter(cell_type == type) %>%
      pull(gencode_id)
   sub <- sanitized.dds[id]
   tpm_score <- colMeans(log10(assays(sub)[["TPM"]]+1))
})

celltype_tpm <- as.data.frame(celltype_tpm) %>%
  rownames_to_column(var="sample_name") %>%
  add_column(pheno_type = sanitized.dds$pheno_type,
             classes = sanitized.dds$new_cluster_name,
             classes_color = sanitized.dds$cluster_color)

9.3 Visualization

The blood, fat and muscle content scores are given to each sample. Boxplot below illustrates that the FSHD biopsies categorized in the Muscle-Low class are fatter and “blooder” and have less muscle content than the rest of the biopsies.

tmp <- celltype_tpm %>% gather(key=cell_type, value=score, -sample_name,
                               -pheno_type, -classes, -classes_color) %>%
  dplyr::mutate(cell_type = factor(cell_type))

color_manual <- celltype_tpm %>% group_by(classes) %>% 
  summarise(color=unique(classes_color)) 

ggplot(tmp, aes(x=classes, y=score, color=classes)) +
  geom_boxplot() +
  facet_wrap(~ cell_type, scale="free",  nrow=2) +
  theme_minimal() +
  theme(axis.text.x=element_blank(), legend.justification=c(1,0), legend.position=c(1,0)) +
  scale_color_manual(values=color_manual$color) +
  labs(y="Avg. log10(TPM +1)", x="FSHD classes", 
       title="Blood, fat, muscle content score by TPM")

9.4 Determine “outliers”

use permutation test? calculate the p-value by constructing mean distribution?

9.5 Determin “bad” biopsies?

9.6 Tables

tb <- dplyr::select(celltype_tpm, -classes_color) %>%
  arrange(classes)
knitr::kable(tb, caption="Blood, fat and muscle content index.")
(#tab:kable_celltype_tpm)Blood, fat and muscle content index.
sample_name blood fat muscle pheno_type classes
01-0041 2.585593 0.1184236 3.862732 Control Control
01-0042 2.604758 0.3822024 3.404674 Control Control
01-0043 1.515110 0.2021819 3.643485 Control Control
01-0044 2.256298 0.2361598 3.655246 Control Control
01-0045 2.485550 0.3499366 3.707235 Control Control
01-0046 3.001476 0.2935164 3.427649 Control Control
01-0048 2.154881 0.3454669 3.692780 Control Control
01-0049 2.199308 0.3796754 3.562354 Control Control
01-0023 1.668277 0.5237397 3.177025 FSHD Mild
01-0032 3.282519 0.5813589 3.790697 FSHD Mild
01-0033 2.471344 0.2054422 3.418840 FSHD Mild
01-0034 2.912924 0.2439875 3.968929 FSHD Mild
01-0036 2.292336 0.3440663 3.634696 FSHD Mild
32-0011 3.213565 0.2619783 3.981536 FSHD Mild
32-0012 3.013254 0.6171513 3.480103 FSHD Mild
32-0014 4.299332 0.1537942 3.158579 FSHD Mild
32-0015 2.891994 0.3012466 4.061947 FSHD Mild
32-0018 3.617742 0.1898964 3.597990 FSHD Mild
32-0019 2.280754 0.3573577 3.984106 FSHD Mild
01-0023b 2.273485 0.7225922 3.462251 FSHD Mild
01-0026b 3.076727 0.3619324 3.636448 FSHD Mild
01-0036b 2.757293 0.3583505 3.653238 FSHD Mild
32-0013b 2.824210 0.6457411 3.511864 FSHD Mild
32-0015b 2.741116 0.3704249 4.023872 FSHD Mild
32-0017b 3.553026 0.3916796 3.089687 FSHD Mild
01-0021 2.883033 0.4052833 3.240677 FSHD Moderate
01-0022-1 2.603183 1.5294705 3.530700 FSHD Moderate
01-0024 3.011862 0.4650832 3.470005 FSHD Moderate
01-0026 2.393720 0.4211012 3.681180 FSHD Moderate
01-0030 3.123735 1.2808079 3.732009 FSHD Moderate
01-0035 2.397119 0.9189982 3.448796 FSHD Moderate
01-0038 2.478129 0.3810781 3.103982 FSHD Moderate
32-0002 2.912887 0.5254690 3.199493 FSHD Moderate
32-0007 2.481612 0.9001335 3.186218 FSHD Moderate
32-0008 3.944364 1.7901614 2.653674 FSHD Moderate
32-0010 3.251255 0.2792022 4.161168 FSHD Moderate
32-0013 3.600465 0.9677029 3.692289 FSHD Moderate
01-0022b 3.330682 1.7365568 3.796104 FSHD Moderate
01-0024b 3.183899 0.2378188 3.327007 FSHD Moderate
01-0028b 2.163574 0.5596671 3.756815 FSHD Moderate
01-0030b 2.371968 0.6290769 3.702877 FSHD Moderate
01-0033b 1.301926 0.5558120 3.672376 FSHD Moderate
01-0034b 2.695132 0.5194755 4.035298 FSHD Moderate
01-0035b 2.638491 1.0832977 3.332846 FSHD Moderate
32-0006b 3.557061 1.0401034 3.582340 FSHD Moderate
32-0009b 2.429373 1.5374834 3.066060 FSHD Moderate
32-0014b 3.373019 0.8191356 3.226881 FSHD Moderate
01-0027 2.861061 1.4552309 3.341357 FSHD IG-High
32-0017 3.645956 1.6709002 3.011805 FSHD IG-High
01-0025b 3.760612 1.5639605 4.035969 FSHD IG-High
01-0027b 3.458966 1.8465812 3.437264 FSHD IG-High
01-0029b 4.002951 1.8184439 3.068390 FSHD IG-High
32-0007b 2.808018 1.7246814 3.398351 FSHD IG-High
32-0018b 3.137937 1.1417576 2.966068 FSHD IG-High
01-0025 2.279839 0.6538198 3.906637 FSHD High
01-0029 3.213923 1.1415584 3.201525 FSHD High
32-0003 2.920859 1.1202031 3.604562 FSHD High
32-0004 2.000568 1.2400798 3.354000 FSHD High
32-0005 4.325426 1.2125240 3.661033 FSHD High
32-0006 2.389504 1.1906286 3.322845 FSHD High
32-0009 3.518911 2.3004965 2.868475 FSHD High
32-0002b1 3.798370 0.6701894 3.311355 FSHD High
32-0005b 4.165111 0.8942373 4.057378 FSHD High
01-0037 3.820057 2.5362662 1.146418 FSHD Muscle-Low
32-0016 4.296978 2.5212618 1.444503 FSHD Muscle-Low
01-0037b 3.680353 1.3811564 3.556671 FSHD Muscle-Low
32-0010b 4.148983 2.2940258 1.548513 FSHD Muscle-Low
32-0012b 4.738981 2.6884333 1.356221 FSHD Muscle-Low
32-0016b 4.944297 0.8406186 1.415840 FSHD Muscle-Low 
#' blood, fat and muscle mean score for each class
tb2 <- tb %>%
  group_by(classes) %>%
  summarise(blood_mean = mean(blood),
            fat_mean = mean(fat),
            muscle_mean = mean(muscle))
knitr::kable(tb2)
classes blood_mean fat_mean muscle_mean
Control 2.350372 0.2884454 3.619519
Mild 2.892347 0.3900435 3.625400
Moderate 2.823931 0.8446781 3.481763
IG-High 3.382214 1.6030794 3.322743
High 3.179168 1.1581930 3.476423
Muscle-Low 4.271608 2.0436270 1.744694