Read in data

## Parsed with column specification:
## cols(
##   sample = col_character(),
##   donor = col_character(),
##   tissue = col_character(),
##   sex = col_character(),
##   age_at_death = col_character(),
##   age_of_onset = col_double(),
##   disease_group = col_character(),
##   site_of_motor_onset = col_character(),
##   reported_mutations = col_character(),
##   RIN = col_double(),
##   sequencing_platform = col_character(),
##   library_size = col_double(),
##   STMN2_TPM = col_double(),
##   tSTMN2_counts = col_double(),
##   tSTMN2_TPM = col_double()
## )
## [1] 0.2399036

Full length STMN2 expression

stmn2_total <-
  stmn2_data %>%
  filter( disease %in% c("Control", "ALS-TDP")) %>%
  mutate( disease = forcats::fct_relevel(disease, "Control")) %>%
  filter( !is.na(sequencing_platform)) %>%
  filter( !is.na(RIN)) %>%
filter(tissue %in% c("Frontal Cortex", "Cerebellum", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Lateral Motor Cortex", "Medial Motor Cortex")) 


stmn2_full_expression_plot <- 
  stmn2_total %>%
  ggplot(aes( x = disease, y = log2(STMN2_TPM + 1), group = disease )) + 
  geom_point(aes(colour = disease),size = 0.5, position = position_jitter(width = 0.2, height = 0) ) + 
  facet_wrap(~sequencing_platform + tissue, ncol = 6) +
  labs(y = expression(log[2]~(TPM+1)), x = "", title = "STMN2 total gene expression by disease and library prep") +
  scale_colour_manual(values = c("gray48", "firebrick")) +
  guides(colour = FALSE) +
  theme_jh() +
  #ylim(0,10 ) +
  geom_boxplot(fill = NA, outlier.colour = NA) +
  theme(panel.border = element_rect(fill =NA)) +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "centre", method = "wilcox.test" )

stmn2_full_RIN_boxplot <- 
  stmn2_total %>%
  #filter( sequencing_platform %in% c("HiSeq 2500", "NovaSeq")) %>%
  ggplot(aes( x = disease, y = RIN )) + 
  geom_point(aes(colour = disease), size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) + 
  facet_wrap(~ sequencing_platform + tissue, ncol = 6) +
  #ylim(2,10) +
  labs(y = "RIN", x = "", title = "Effect of sequencing sequencing_platform on RIN in ALS-TDP samples") +
  scale_colour_manual(values = c("gray48", "firebrick")) +
  scale_alpha_manual(values = c(1,0.1) ) +
  guides(colour = FALSE) +
  theme_jh() +
  geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  theme(panel.border = element_rect(fill =NA)) +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "centre", method = "wilcox.test")


stmn2_full_RIN_corplot <- 
  stmn2_total %>%
  ggplot(aes( x = RIN, y = log2(STMN2_TPM + 1) )) + 
  geom_point(aes(colour = disease), size = 0.5, alpha = 0.9) +
  scale_colour_manual(values = c("gray48", "firebrick")) +
  facet_wrap(~ sequencing_platform + tissue, ncol = 6) +
  labs(y = expression(log[2]~(TPM+1)), x = "RIN", title = "STMN2 total gene expression and RIN") +
  scale_alpha_manual(values = c(1,0.1)) +
  #ylim(0,10) + 
  guides(colour = FALSE) +
  theme_jh() +
  theme(panel.border = element_rect(fill =NA)) +
  geom_smooth(method = "lm", colour = "black") + 
  ggpubr::stat_cor(colour = "black",method = "spearman")


stmn2_full_expression_plot 

## `geom_smooth()` using formula 'y ~ x'

Truncated STMN2 novel splicing in full dataset

Treat as binary outcome

Plot all junction discovery rates in each condition and tissue

tstmn2_full_dataset <- 
  stmn2_data %>% 
  mutate(stmn2_detection = tSTMN2_counts > 1) %>%
  group_by(disease) %>%
  summarise( samples = n(), stmn2_n = sum(stmn2_detection), stmn2_prop = sum(stmn2_detection) / n() ) %>%
  mutate( stmn2_prop_label = paste0(round(stmn2_prop * 100, digits = 1), "%")) %>%
  mutate( stmn2_prop_label = ifelse( stmn2_prop_label == "0%", "", stmn2_prop_label)) %>%
 # mutate( stmn2_prop_label = gsub("\\.", "·", stmn2_prop_label)) %>%
  mutate(tissue = "Full Dataset")

tstmn2_by_tissue <- 
  stmn2_data %>%
  mutate(stmn2_detection = tSTMN2_counts > 1) %>%
  group_by(tissue, disease) %>%
  summarise( samples = n(), stmn2_n = sum(stmn2_detection), stmn2_prop = sum(stmn2_detection) / n() ) %>%
  mutate( stmn2_prop_label = paste0(round(stmn2_prop * 100, digits = 1), "%")) %>%
  mutate( stmn2_prop_label = ifelse( stmn2_prop_label == "0%", "", stmn2_prop_label)) %>%
 # mutate( stmn2_prop_label = gsub("\\.", "·", stmn2_prop_label)) %>%
  ungroup() 


tstmn2_all_samples <- 
  tibble(tissue = "Full Dataset", disease = "All samples", samples = nrow(stmn2_data), stmn2_n = stmn2_data %>% filter(tSTMN2_counts > 1) %>% nrow() ) %>%
  mutate(stmn2_prop = stmn2_n / samples) %>%
  mutate(stmn2_prop_label = paste0( round(stmn2_prop * 100, digits = 1), "%")) #%>%
  #mutate( stmn2_prop_label = gsub("\\.", "·", stmn2_prop_label))


stmn2_to_plot <-
  bind_rows(
    tstmn2_by_tissue,
    tstmn2_full_dataset,
    tstmn2_all_samples
  ) %>%
   mutate(tissue = factor(tissue, levels = c("Full Dataset", "Frontal Cortex", "Temporal Cortex", "Hippocampus", "Lateral Motor Cortex", "Medial Motor Cortex", "Other Motor Cortex", "Occipital Cortex", "Cerebellum",   "Cervical Spinal Cord", "Thoracic Spinal Cord", "Lumbar Spinal Cord") )) %>%
  mutate(disease = factor(disease, levels = c("All samples", "Control", "ALS-TDP", "ALS/AD", "ALS-SOD1", "ALS/FTLD", "FTLD-TDP", "FTLD-FUS", "FTLD-TAU"))) %>%
  filter( !is.na(tissue) & !is.na(disease)) %>%
  mutate(disease = forcats::fct_rev(disease))
## Warning in bind_rows_(x, .id): binding factor and character vector,
## coercing into character vector
## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector
## Warning in bind_rows_(x, .id): binding factor and character vector,
## coercing into character vector
## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector
stmn2_threshold_plot <-
  stmn2_to_plot %>%
  ggplot(aes(x = disease, y = stmn2_prop ) ) +
  geom_col(aes(fill = disease), colour = "white") +
  #facet_wrap(~Tissue) +
  #coord_flip() + 
  geom_text(data = select(stmn2_to_plot, tissue, disease, samples) %>% distinct(), aes(label = samples), y = 1.1, size = 3.5) +
  geom_text(data = filter(stmn2_to_plot, stmn2_prop < 0.8), aes(label = stmn2_prop_label) , nudge_y = 0.1, size = 3.5) +
  geom_text(data = filter(stmn2_to_plot, stmn2_prop >= 0.8), aes(label = stmn2_prop_label), nudge_y = -0.12, size = 3.5, colour = "white") +
  scale_y_continuous(expand = c(0,0), limits = c(0,1.2),labels = scales::percent, breaks = c(0,0.25,0.5,0.75,1) ) +
  #geom_errorbar(aes(ymin = prop, ymax = binom_conf_upper)) +
  facet_wrap(~tissue, scales = "free_y") +
  geom_hline(yintercept = 1, linetype = 3) +
  labs(x = "", y = "Truncated STMN2 detection rate") + 
  coord_flip() +
  #labs(y = "% of samples with 2 or more novel STMN2 junction reads", x = "", title = "STMN2 novel splicing across disease and tissue") +
  theme_jh() +
  scale_fill_manual(values = c("Control" = "gray", "ALS-TDP" = "firebrick", "FTLD-TDP" = "darkblue", "ALS/FTLD" = "purple4", "ALS/AD" = "maroon", "ALS-SOD1" = "goldenrod", "FTLD-TAU" = "darkblue", "FTLD-FUS" = "darkblue", "All samples" = "black")) +
  #scale_alpha_manual("read\nthreshold",values = c(0.3,0.6,1) ) +
  theme( panel.border = element_rect(fill = NA), axis.line = element_blank(), plot.title = element_text(size = 14), 
         strip.text = element_text(face = "bold", margin = margin(t = 5, r = 0, b = 2, l = 0, unit = "pt")), plot.background = element_rect(fill = "white", colour = NA) ) +
  guides(fill = FALSE)

stmn2_threshold_plot

Selected Truncated STMN2 detection rates

ALS and FTLD

# add sample numbers to condition labels
stmn2_main_plot <-
  tstmn2_by_tissue %>%
  mutate(tissue = gsub(" ", "\n", tissue)) %>%
  mutate(disease = factor(disease, levels = c("Control", "ALS-TDP", "ALS/FTLD", "ALS/AD", "FTLD-TDP", "ALS-SOD1", "FTLD-FUS", "FTLD-TAU"))) %>%
  mutate(tissue = factor(tissue, levels = c("Frontal\nCortex","Lateral\nMotor\nCortex", "Medial\nMotor\nCortex", "Other\nMotor\nCortex", "Temporal\nCortex", "Hippocampus", "Occipital\nCortex", "Cerebellum", "Cervical\nSpinal\nCord", "Thoracic\nSpinal\nCord", "Lumbar\nSpinal\nCord")) ) %>%
  mutate( sample_n = paste0(disease,"\nN = ",samples)) %>%
  mutate(tissue_n = paste0(tissue, "\nN = ",samples) ) %>%
  filter(!is.na(tissue))

# all tissues, FTLD-TDP
FTLD_all <- 
  stmn2_main_plot %>%
  filter(disease %in% c("FTLD-TDP") )  %>%
  arrange( tissue) %>%
  mutate( tissue_n = factor(tissue_n,levels = tissue_n) )

FTLD_all_plot <- 
  FTLD_all %>%
  ggplot(aes(x = tissue_n, y = stmn2_prop ) ) +
  geom_col(fill = "darkblue", colour = "white") +
  scale_y_continuous(expand = c(0,0), limits = c(0,1.1),labels = scales::percent ) +
  #geom_text(data = select(., tissue, disease, samples) %>% distinct(), aes(label = samples), y = 1.05, size = 3) #+
  geom_text(data = filter(FTLD_all, stmn2_prop < 0.9), aes(label = stmn2_prop_label) , nudge_y = 0.05, size = 3) +
  geom_text(data = filter(FTLD_all, stmn2_prop >= 0.9), aes(label = stmn2_prop_label), nudge_y = -0.05, size = 3, colour = "white") +
  theme_jh() +
  labs(title = "FTLD-TDP by tissue", y = "Truncated STMN2 detection", x = "") +
  theme(plot.title = element_text(hjust = 0.5))

als_all <- 
  stmn2_main_plot %>%
  filter(disease %in% c("ALS-TDP") )%>%
  #filter( tissue %in%  c("Frontal\nCortex", "Lateral\nMotor\nCortex", "Medial\nMotor\nCortex", "Occipital\nCortex", "Cerebellum", "Cervical\nSpinal\nCord", "Lumbar\nSpinal\nCord")) %>%
  arrange( tissue) %>%
  mutate( tissue_n = factor(tissue_n,levels = tissue_n) )

als_all_plot <- 
  als_all %>%
  ggplot(aes(x = tissue_n, y = stmn2_prop ) ) +
  geom_col(fill = "firebrick", colour = "white") +
  scale_y_continuous(expand = c(0,0), limits = c(0,1.1),labels = scales::percent ) +
  #geom_text(data = select(., tissue, disease, samples) %>% distinct(), aes(label = samples), y = 1.05, size = 3) #+
  geom_text(data = filter(als_all, stmn2_prop < 0.9), aes(label = stmn2_prop_label) , nudge_y = 0.05, size = 3) +
  geom_text(data = filter(als_all, stmn2_prop >= 0.9), aes(label = stmn2_prop_label), nudge_y = -0.05, size = 3, colour = "white") +
  theme_jh() +
  labs(title = "ALS-TDP by tissue", y = "Truncated STMN2 detection", x = "") +
  theme(plot.title = element_text(hjust = 0.5))


als_matrix <- stmn2_data %>% 
  mutate(disease = factor(disease, levels = c("Control", "ALS-TDP", "FTLD-TDP", "ALS-SOD1", "FTLD-FUS", "FTLD-TAU", "ALS/FTLD"))) %>%
  #filter(disease == "ALS-TDP") %>%
  mutate(stmn = as.numeric(tSTMN2_counts > 1) ) %>%
  filter( disease %in% c("Control", "ALS-TDP", "ALS-SOD1")) %>%
  filter(tissue %in%  "Lumbar Spinal Cord") %>%
  select(donor, tissue, stmn, disease) %>%
  group_by(donor, disease) %>%
  summarise( n_tissues = sum(stmn, na.rm = TRUE)) %>%
  mutate( stmn_hit = n_tissues > 0) %>%
  group_by( disease) %>%
  summarise( samples = n() , stmn_prop = sum(stmn_hit) / n() ) %>%
  mutate( stmn2_prop_label = paste0(signif(stmn_prop * 100, digits = 3), "%" ) ) %>%
  #mutate( stmn2_prop_label = gsub("\\.", "·", stmn2_prop_label)) %>%
  mutate( stmn2_prop_label = ifelse( stmn2_prop_label == "0%", "", stmn2_prop_label)) %>%
  mutate( sample_n = paste0(disease,"\nN = ",samples)) %>%
  mutate( sample_n = factor(sample_n,levels = sample_n) )

als_total_plot <- 
  als_matrix %>%
  ggplot(aes(x = sample_n, y = stmn_prop ) ) +
  geom_col(fill = "firebrick", colour = "white") +
  scale_y_continuous(expand = c(0,0), limits = c(0,1),labels = scales::percent ) +
  geom_text(data = filter(als_matrix, stmn_prop < 0.9), aes(label = stmn2_prop_label) , nudge_y = 0.05, size = 3) +
  geom_text(data = filter(als_matrix, stmn_prop >= 0.9), aes(label = stmn2_prop_label), nudge_y = -0.05, size = 3, colour = "white") +
  labs(title = "ALS-TDP\nLumbar Spinal Cord", x = "", y = "Truncated STMN2 detection") +
  theme_jh() +
  theme(plot.title = element_text(hjust = 0.5))


FTLD_matrix <- 
  stmn2_data %>% 
  mutate(disease = factor(disease, levels = c("Control", "ALS-TDP", "FTLD-TDP", "ALS-SOD1", "FTLD-FUS", "FTLD-TAU", "ALS/FTLD"))) %>%
  filter( disease %in% c("Control", "FTLD-TDP", "FTLD-TAU", "FTLD-FUS")) %>%
  mutate(stmn = as.numeric(tSTMN2_counts > 1) ) %>%
  filter(tissue %in% c("Frontal Cortex")) %>%
  select(donor, tissue, stmn, disease) %>%
  group_by(donor, disease) %>%
  summarise( n_tissues = sum(stmn, na.rm = TRUE)) %>%
  mutate( stmn_hit = n_tissues > 0) %>%
  group_by( disease) %>%
  summarise( samples = n() , stmn_prop = sum(stmn_hit) / n() ) %>%
  mutate( stmn2_prop_label = paste0(signif(stmn_prop * 100, digits = 3), "%" ) ) %>%
  mutate( stmn2_prop_label = ifelse( stmn2_prop_label == "0%", "", stmn2_prop_label)) %>%
  #mutate( stmn2_prop_label = gsub("\\.", "·", stmn2_prop_label)) %>%
  mutate( sample_n = paste0(disease,"\nN = ",samples)) %>%
  mutate( sample_n = factor(sample_n,levels = sample_n) )


FTLD_total_plot <- 
  FTLD_matrix %>%
  ggplot(aes(x = sample_n, y = stmn_prop ) ) +
  geom_col(fill = "darkblue", colour = "white") +
  scale_y_continuous(expand = c(0,0), limits = c(0,1),labels = scales::percent ) +
  geom_text(data = filter(FTLD_matrix, stmn_prop < 0.9), aes(label = stmn2_prop_label) , nudge_y = 0.05, size = 3) +
  geom_text(data = filter(FTLD_matrix, stmn_prop >= 0.9), aes(label = stmn2_prop_label), nudge_y = -0.05, size = 3, colour = "white") +
  labs(title = "FTLD-TDP\nFrontal Cortex", x = "", y = "Truncated STMN2 detection") +
  theme_jh() +
  theme(plot.title = element_text(hjust = 0.5))

#library(magick)
#schematic_img <- image_read("../figures/STMN2_schematic.png" )

schematic <- ggplot() + theme_void()

tSTMN2_detection_spotlight <- 
  (schematic + FTLD_total_plot + als_total_plot  ) /
(FTLD_all_plot + als_all_plot + plot_layout(widths = c(1.8,4) ) )  +
  plot_annotation(tag_levels = 'A') &
  theme(plot.tag = element_text(size = 14, face = "bold" ), plot.title = element_text(face = "bold"), strip.background = element_blank(), plot.background = element_blank(),panel.background = element_blank(), panel.border = element_blank() )

tSTMN2_detection_spotlight

ALS/FTLD and ALS/AD

als_ftd_all <- 
  stmn2_main_plot %>%
  filter(disease %in% c("ALS/FTLD") )%>%
  filter( tissue %in%  c("Frontal\nCortex", "Hippocampus", "Lateral\nMotor\nCortex", "Medial\nMotor\nCortex", "Occipital\nCortex", "Cerebellum", "Cervical\nSpinal\nCord", "Lumbar\nSpinal\nCord")) %>%
  arrange( tissue) %>%
  mutate( tissue_n = factor(tissue_n,levels = tissue_n) )

als_ftd_all_plot <- 
  als_ftd_all %>%
  ggplot(aes(x = tissue_n, y = stmn2_prop ) ) +
  geom_col(fill = "purple4", colour = "white") +
  scale_y_continuous(expand = c(0,0), limits = c(0,1.1),labels = scales::percent ) +
  #geom_text(data = select(., tissue, disease, samples) %>% distinct(), aes(label = samples), y = 1.05, size = 3) #+
  geom_text(data = filter(als_ftd_all, stmn2_prop < 0.9), aes(label = stmn2_prop_label) , nudge_y = 0.05, size = 3) +
  geom_text(data = filter(als_ftd_all, stmn2_prop >= 0.9), aes(label = stmn2_prop_label), nudge_y = -0.05, size = 3, colour = "white") +
  theme_jh() +
  labs(title = "ALS/FTLD by tissue", y = "Truncated STMN2 detection", x = "") +
  theme(plot.title = element_text(hjust = 0.5))


als_ad_all <- 
  stmn2_main_plot %>%
  filter(disease %in% c("ALS/AD") )%>%
  filter( tissue %in%  c("Frontal\nCortex", "Hippocampus", "Lateral\nMotor\nCortex", "Medial\nMotor\nCortex", "Occipital\nCortex", "Cerebellum", "Cervical\nSpinal\nCord", "Lumbar\nSpinal\nCord")) %>%
  arrange( tissue) %>%
  mutate( tissue_n = factor(tissue_n,levels = tissue_n) )

als_ad_all_plot <- 
  als_ad_all %>%
  ggplot(aes(x = tissue_n, y = stmn2_prop ) ) +
  geom_col(fill = "maroon", colour = "white") +
  scale_y_continuous(expand = c(0,0), limits = c(0,1.1),labels = scales::percent ) +
  #geom_text(data = select(., tissue, disease, samples) %>% distinct(), aes(label = samples), y = 1.05, size = 3) #+
  geom_text(data = filter(als_ad_all, stmn2_prop < 0.9), aes(label = stmn2_prop_label) , nudge_y = 0.05, size = 3) +
  geom_text(data = filter(als_ad_all, stmn2_prop >= 0.9), aes(label = stmn2_prop_label), nudge_y = -0.05, size = 3, colour = "white") +
  theme_jh() +
  labs(title = "ALS/AD by tissue", y = "Truncated STMN2 detection", x = "") +
  theme(plot.title = element_text(hjust = 0.5))

als_ftd_als_ad_plot <- 
  als_ftd_all_plot +
  als_ad_all_plot +
  plot_layout(nrow = 2)  +plot_annotation(tag_levels = 'A') &
  theme(plot.tag = element_text(size = 14, face = "bold" ) )


als_ftd_als_ad_plot

Truncated STMN2 expression levels

# diagnostic plot
tstmn2_diagnostics <- 
  stmn2_data %>%
   mutate( disease = factor(disease, levels = c("Control", "ALS-TDP", "FTLD-TDP", "ALS/FTLD", "ALS/AD"))) %>%
  filter(!is.na(sequencing_platform)) %>%
  filter(tSTMN2_counts > 1) %>%
  filter( disease %in% c("ALS-TDP", "FTLD-TDP", "ALS/FTLD", "ALS/AD")) %>%
  filter(tissue %in% c("Frontal Cortex", "Temporal Cortex", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Lateral Motor Cortex", "Medial Motor Cortex")) 

# raw counts
counts_by_disease_plot <-  
  tstmn2_diagnostics %>%
  ggplot(aes( x = disease, y = tSTMN2_counts, group = disease, colour = disease )) + 
  geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  geom_point(size = 0.5, position = position_jitter(width = 0.2, height = 0) ) + 
  scale_y_continuous(trans='log10', breaks = c(2, 10,100), limits = c(2, 130)) +
  facet_wrap(~tissue, nrow = 1, scales = "free_x") +
  scale_colour_manual(values = c("Control" = "gray", "ALS-TDP" = "firebrick", "FTLD-TDP" = "darkblue", "ALS/FTLD" = "purple4", "ALS/AD" = "maroon")) +
  labs(y = "raw counts", x = "", subtitle = "") +
  guides(colour = FALSE) +
  theme_jh() +
  theme(panel.border = element_rect(fill =NA))

# counts per millino
tpm_by_disease_plot <-
  tstmn2_diagnostics %>%
  ggplot(aes( x = disease, y = tSTMN2_TPM, group = disease, colour = disease )) + 
  geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  geom_point(size = 0.5, position = position_jitter(width = 0.2, height = 0) ) + 
    scale_y_continuous(trans='log10', breaks = c(0.1,1), limits = c(0.01, 2.5)) +
  facet_wrap(~tissue, nrow = 1, scales = "free_x") +
  scale_colour_manual(values = c("Control" = "gray", "ALS-TDP" = "firebrick", "FTLD-TDP" = "darkblue", "ALS/FTLD" = "purple4", "ALS/AD" = "maroon")) +
  labs(y = "counts per million", x = "") +
    labs(y = "counts per million", x = "", subtitle = "") +
  guides(colour = FALSE) +
  #ylim(0,1.8) + 
  theme_jh() +
  theme(panel.border = element_rect(fill =NA))


tstmn2_counts_plot <- 
  counts_by_disease_plot +
  tpm_by_disease_plot +
  plot_annotation(tag_levels = "A") +
  plot_layout(nrow = 2) & 
  theme(strip.text = element_text(margin = margin(t = 1, r = 0, b = 5, l = 0, unit = "pt")),
        plot.tag = element_text(face = "bold",size = 14), 
        panel.spacing = unit(1, "lines")  ) & coord_flip()

tstmn2_counts_plot

Effect of sequencing_platform and/or RIN on Truncated STMN2 detection in ALS-TDP

## effect of RIN?
als_RIN_plot <- 
  stmn2_data %>%
  filter(!is.na(RIN)) %>%
  filter(disease %in% c("ALS-TDP")) %>%
  filter( tissue %in% c("Frontal Cortex", "Thoracic Spinal Cord", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Medial Motor Cortex", "Lateral Motor Cortex")) %>%
  mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") ) %>% 
  ggplot(aes(x = stmn2_data,y = RIN )) + 
  geom_point( colour = "firebrick", size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) + 
    geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  labs(title = "Truncated STMN2 detection and RIN") + 
  facet_wrap(~disease + tissue, ncol = 6 ) + theme_jh() +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "center" ) +
  labs(y = "RIN", x = "Truncated STMN2 detection") +
  theme(axis.text.x = element_text(size = 15))

stmn2_sequencing_platform <- 
  stmn2_data %>%
  filter(!is.na(RIN)) %>%
  filter( tissue %in% c("Frontal Cortex","Thoracic Spinal Cord", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Medial Motor Cortex", "Lateral Motor Cortex")) %>%
  mutate(tissue = as.character(tissue)) %>%
  filter( !is.na(sequencing_platform)) %>%
  filter(disease %in% c("ALS-TDP")) %>%
    mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") )

stmn2_data_fisher <- 
  split(stmn2_sequencing_platform, stmn2_sequencing_platform$tissue) %>%
  purrr::map_df( ~{tibble( p_value = fisher.test(table(stmn2 = .x$stmn2_data, sequencing_platform = .x$sequencing_platform  ))$p.value) }, .id = "tissue" ) 

stmn2_sequencing_platform_pct <-
  stmn2_sequencing_platform %>%
  group_by(tissue, sequencing_platform, stmn2_data) %>% tally() %>%
  spread(key = stmn2_data, value = n) %>%
  mutate(detection = `+` / ( `-` + `+`) )



als_sequencing_platform_plot <- 
  stmn2_sequencing_platform_pct %>%
  ggplot(aes(x = sequencing_platform, y = detection )) + 
  geom_col(fill = "firebrick") +
  labs(title = "Truncated STMN2 detection and sequencing platform", y = "Truncated STMN2 detection") + 
  facet_wrap(~tissue, ncol = 6 ) +
  theme_jh() +
  scale_y_continuous(labels = scales::percent_format(accuracy = 1)) +
  #scale_fill_manual("STMN2\nnovel\njunction", values = c("white", "firebrick")) +
  geom_text(data = stmn2_data_fisher, aes(x = 1.5, y = 1.05, label = paste0("p = ", signif(p_value,digits = 2) ) )) +
  theme(legend.text = element_text(size = 15))
 # scale_y_continuous(expand = c(0,0) ) 

als_RIN_plot + als_sequencing_platform_plot + plot_layout(nrow = 2)

Effect of library size and full length STMN2 expression on truncated STMN2 discovery in ALS-TDP

# library size 

als_plot_a <- 
  stmn2_data %>%
  filter(disease %in% c("ALS-TDP")) %>%
  filter( tissue %in% c("Frontal Cortex", "Thoracic Spinal Cord", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Medial Motor Cortex", "Lateral Motor Cortex")) %>%
  mutate( tissue = factor( tissue, levels = c("Frontal Cortex", "Thoracic Spinal Cord", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Medial Motor Cortex", "Lateral Motor Cortex")) ) %>%
  mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") ) %>% 
  ggplot(aes(x = stmn2_data,y = log10(library_size) )) + 
  geom_point(colour = "firebrick", size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) + 
    geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  labs(title = "Truncated STMN2 detection and library size") + 
  facet_wrap(~ tissue, ncol = 6 ) + theme_jh() +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "center", method = "wilcox" ) +
  labs(y = "log10 library size", x = "Truncated STMN2 detection") +
  theme(axis.text.x = element_text(size = 15))


als_plot_d <- 
  stmn2_data %>%
    filter(disease %in% c("ALS-TDP")) %>%
  filter(!is.na(sequencing_platform)) %>%
  filter( tissue %in% c("Frontal Cortex", "Thoracic Spinal Cord", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Medial Motor Cortex", "Lateral Motor Cortex")) %>%
  mutate( tissue = factor( tissue, levels = c("Frontal Cortex", "Thoracic Spinal Cord", "Cervical Spinal Cord", "Lumbar Spinal Cord", "Medial Motor Cortex", "Lateral Motor Cortex")) ) %>%
  mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") ) %>% 
  ggplot(aes(x = stmn2_data,y = STMN2_TPM )) + 
  geom_point( colour = "firebrick", size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) + 
  geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  labs(title = "Truncated STMN2 detection and STMN2 expression ") + 
  facet_wrap(~ tissue, ncol = 6,scales = "free_y") + theme_jh() +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "center", method = "wilcox" ) +
  labs(y = "STMN2 expression (TPM)", x = "Truncated STMN2 detection") +
  theme(axis.text.x = element_text(size = 15))


stmn2_data_technical_plot <- 
 als_plot_a + als_plot_d + plot_layout(nrow = 2)

stmn2_data_technical_plot

Technical factors and STMN2 full-length expression explaining tSTMN2 in FTD-TDP

# library size 

ftd_plot_libsize <- 
  stmn2_data %>%
  filter(disease %in% c("FTLD-TDP")) %>%
  filter( tissue %in% c("Frontal Cortex",  "Temporal Cortex" )) %>%
  mutate(tissue = factor(tissue, levels = c("Frontal Cortex", "Temporal Cortex"))) %>%
  mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") ) %>% 
  ggplot(aes(x = stmn2_data,y = log10(library_size) )) + 
  geom_point( colour = "darkblue", size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) + 
    geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  #labs(title = "Truncated STMN2 detection and library size", subtitle = "FTD-TDP samples") + 
  facet_wrap(~ tissue, ncol = 6 ) + theme_jh() +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "center", method = "wilcox" ) +
  labs(y = "log10 library size", x = "Truncated STMN2 detection") +
  theme(axis.text.x = element_text(size = 15))

# percent coding bases
# 
# ftd_plot_coding_bases <- 
#   stmn2_data %>%
#     left_join(tech_df) %>%
#   filter(disease %in% c("FTLD-TDP")) %>%
#   filter( tissue %in% c("Frontal Cortex",  "Temporal Cortex" )) %>%
#   mutate(tissue = factor(tissue, levels = c("Frontal Cortex", "Temporal Cortex", "Cerebellum" ))) %>%
#   mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") ) %>% 
#   ggplot(aes(x = stmn2_data,y = pct_coding_bases )) + 
#   geom_point( colour = "darkblue",size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) + 
#     geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
#   #labs(title = "Truncated STMN2 detection and % coding bases ", subtitle = "FTD-TDP samples") + 
#   facet_wrap(~  tissue, ncol = 6 ) + theme_jh() +
#   ggpubr::stat_compare_means(label = "p.format", label.x.npc = "center", method = "wilcox" ) +
#   labs(y = "% coding bases", x = "Truncated STMN2 detection") +
#   theme(axis.text.x = element_text(size = 15))
# 
# # coverage bias
# 
ftd_plot_RIN <-
  stmn2_data %>%
  filter(disease %in% c("FTLD-TDP")) %>%
  filter( tissue %in% c("Frontal Cortex",  "Temporal Cortex" )) %>%
  mutate(tissue = factor(tissue, levels = c("Frontal Cortex", "Temporal Cortex", "Cerebellum" ))) %>%
  mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") ) %>%
  ggplot(aes(x = stmn2_data,y = RIN )) +
  geom_point(colour = "darkblue", size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) +
    geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  #labs(title = "Truncated STMN2 detection and 3' coverage bias", subtitle = "FTD-TDP samples") +
  facet_wrap(~ tissue, ncol = 6 ) + theme_jh() +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "center", method = "wilcox" ) +
  labs(y = "RIN", x = "Truncated STMN2 detection") +
  theme(axis.text.x = element_text(size = 15))


ftd_plot_stmn2 <- 
  stmn2_data %>% 
  filter(disease %in% c("FTLD-TDP")) %>%
  filter( tissue %in% c("Frontal Cortex",  "Temporal Cortex" )) %>%
  mutate(tissue = factor(tissue, levels = c("Frontal Cortex", "Temporal Cortex", "Cerebellum" ))) %>%
  mutate(stmn2_data = ifelse(tSTMN2_counts > 1, yes = "+", no = "-") ) %>% 
  ggplot(aes(x = stmn2_data,y = STMN2_TPM )) + 
  geom_point( colour = "darkblue", size = 0.5, alpha = 0.9, position = position_jitter(width = 0.2, height = 0) ) + 
    geom_boxplot(fill = NA, colour = "black", outlier.colour = NA) +
  #labs(title = "Truncated STMN2 detection and STMN2 expression ", subtitle = "ALS-TDP samples") + 
  facet_wrap(~tissue, ncol = 6,scales = "free_y") + theme_jh() +
  ggpubr::stat_compare_means(label = "p.format", label.x.npc = "center", method = "wilcox" ) +
  labs(y = "STMN2 expression (TPM)", x = "Truncated STMN2 detection") +
  theme(axis.text.x = element_text(size = 15))

ftd_technical_plot <- 
  ftd_plot_RIN + ftd_plot_libsize + ftd_plot_stmn2 + plot_annotation(tag_levels = 'A') & theme(plot.tag = element_text(size = 12, face = "bold")) 

#ggsave(plot = ftd_technical_plot, filename = here::here("figures/tSTMN2_truncated_FTD_technical.png"), width =14, height =4, dpi = 300)

ftd_technical_plot

Truncated STMN2 detection in ALS-TDP patients with C9orf72 expansions

Truncated STMN2 detection in FTLD-TDP patients with C9orf72 expansions

## 
##  Fisher's Exact Test for Count Data
## 
## data:  table(stmn2 = ftd_c9$stmn2_data, c9 = ftd_c9$reported_mutations)
## p-value = 1
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.2121109 4.8153877
## sample estimates:
## odds ratio 
##  0.9276436

tSTMN2 detection in ALS-TDP and age of onset and death

In both Cervical and Lumbar Spinal Cord, donors with detectable tSTMN2 have a later age of onset and a shorter length of time with the disease.

What do we make of this? ALS can be a rapidly progressing disease and this is often seen in early onset cases. However lots of early-onset cases have a slowly progressing disease course.

If we are more likely to observe tSTMN2 in cases that had a later disease onset but shorter disease course, what does this mean? Shorter disease course could mean either a more rapid progression or less time for motor neurons to get TDP pathology and die before the patient themselves die, perhaps of complications. We would need more clinical information (staging, measures of decline over time) to properly assess this.

FTLD-TDP - tSTMN2 detection and age at death

Age of disease onset information is not provided for FTLD-TDP samples