Figure 4: Large deletions and genomic integrations
E. coli ORBIT 2023
Scott H. Saunders
Notes
This figure contains several datatypes from experiments where ORBIT was used for large genomic deletions and integrations. Data figures were made in R notebooks and exported as pdfs. Cosmetic improvements were made in Adobe Illustrator. Note that Figures 4A, 4D & 4F were made in Adobe Illustrator.
Setup packages and plotting for the notebook:
# Check packages
source("../tools/package_setup.R")
# Load packages
library(tidyverse)
library(cowplot)
library(kableExtra)
# Code display options
knitr::opts_chunk$set(tidy.opts=list(width.cutoff=60),tidy=FALSE, echo = TRUE, message=FALSE, warning=FALSE, fig.align="center", fig.retina = 2)
# Load plotting tools
source("../tools/plotting_tools.R")
#Modify the plot theme
theme_set(theme_notebook())Fig. 4B - Deletion size efficiency
The galK deletions were done separately from the other loci, so let’s read that data in first.
df_gal_sizes <- read_csv('../../data/low_throughput_experiments/2022_11_09_galK_sizes_3_efficiency.csv') %>% #read in csv
mutate(eff = Kan_count / LB_count) %>% group_by(del_size) %>% mutate(avg_eff = mean(eff)) #calculate efficiency and average efficiency for replicates
#calculate negative control value from galK experiment
gal_sizes_pInt <- (df_gal_sizes %>% filter(del_size == '-'))$avg_eff[1]
df_gal_sizes %>% kable() %>% kable_styling() %>% scroll_box(height = '250px')| locus | del_size | replicate | LB_count | Kan_count | eff | avg_eff |
|---|---|---|---|---|---|---|
| galK | 1122 | 1 | 13900000 | 33000 | 0.0023741 | 0.0026285 |
| galK | 1122 | 2 | 11100000 | 32000 | 0.0028829 | 0.0026285 |
| galK | 4258 | 1 | 10800000 | 1230 | 0.0001139 | 0.0001266 |
| galK | 4258 | 2 | 9900000 | 1380 | 0.0001394 | 0.0001266 |
| galK | 10762 | 1 | 9400000 | 430 | 0.0000457 | 0.0000592 |
| galK | 10762 | 2 | 9500000 | 690 | 0.0000726 | 0.0000592 |
| galK | 24027 | 1 | 10600000 | 990 | 0.0000934 | 0.0000920 |
| galK | 24027 | 2 | 11700000 | 1060 | 0.0000906 | 0.0000920 |
| galK | 49068 | 1 | 11400000 | 160 | 0.0000140 | 0.0000164 |
| galK | 49068 | 2 | 15000000 | 280 | 0.0000187 | 0.0000164 |
| galK |
|
1 | 11900000 | 240 | 0.0000202 | 0.0000243 |
| galK |
|
2 | 8800000 | 250 | 0.0000284 | 0.0000243 |
Now we will read in the data for the other loci (hisA, metA, and leuD).
df_aa_sizes <- read_csv('../../data/low_throughput_experiments/2022_09_07_AA_del_sizes_eff.csv') %>% #read in csv
mutate(eff = Kan_count / LB_count) %>% group_by(locus, del_size) %>% mutate(avg_eff = mean(eff)) #calculate efficiency and average efficiency for replicates
df_aa_sizes %>% kable() %>% kable_styling() %>% scroll_box(height = '250px')| locus | del_size | replicate | LB_count | Kan_count | eff | avg_eff |
|---|---|---|---|---|---|---|
| hisA | 100 | 1 | 370000 | 1120 | 0.0030270 | 0.0024635 |
| hisA | 100 | 2 | 900000 | 1710 | 0.0019000 | 0.0024635 |
| hisA | 700 | 1 | 800000 | 5100 | 0.0063750 | 0.0061420 |
| hisA | 700 | 2 | 880000 | 5200 | 0.0059091 | 0.0061420 |
| hisA | 7000 | 1 | 780000 | 100 | 0.0001282 | 0.0001195 |
| hisA | 7000 | 2 | 830000 | 92 | 0.0001108 | 0.0001195 |
| metA | 100 | 1 | 780000 | 8400 | 0.0107692 | 0.0102826 |
| metA | 100 | 2 | 980000 | 9600 | 0.0097959 | 0.0102826 |
| metA | 900 | 1 | 1250000 | 6000 | 0.0048000 | 0.0044886 |
| metA | 900 | 2 | 790000 | 3300 | 0.0041772 | 0.0044886 |
| metA | 13000 | 1 | 890000 | 1010 | 0.0011348 | 0.0010500 |
| metA | 13000 | 2 | 860000 | 830 | 0.0009651 | 0.0010500 |
| leuD | 100 | 1 | 880000 | 1370 | 0.0015568 | 0.0014955 |
| leuD | 100 | 2 | 760000 | 1090 | 0.0014342 | 0.0014955 |
| leuD | 600 | 1 | 1010000 | 720 | 0.0007129 | 0.0005181 |
| leuD | 600 | 2 | 990000 | 320 | 0.0003232 | 0.0005181 |
| leuD | 6000 | 1 | 760000 | 430 | 0.0005658 | 0.0005980 |
| leuD | 6000 | 2 | 730000 | 460 | 0.0006301 | 0.0005980 |
| galK | 1000 | 1 | 700000 | 3600 | 0.0051429 | 0.0039413 |
| galK | 1000 | 2 | 730000 | 2000 | 0.0027397 | 0.0039413 |
#combine datasets
df_gal_aa_sizes <- bind_rows(df_gal_sizes %>% filter(del_size!='-') %>% mutate(del_size = as.numeric(del_size)),
df_aa_sizes %>% filter(locus!='galK')
) %>%
mutate(locus = factor(locus, levels = c( 'galK','hisA','metA','leuD')))
#Plot with individual datapoints, and mean points connected by lines
plot_gal_aa_sizes <- ggplot(df_gal_aa_sizes, aes(x = del_size, y = eff, color = locus)) +
geom_hline(yintercept = gal_sizes_pInt, linetype = 2, color = 'light gray')+
geom_jitter(shape = 21, width = 0.05, height = 0, alpha = 0.4) +
geom_point(data = . %>% filter(replicate==1), aes(y = avg_eff)) +
geom_line(data = . %>% filter(replicate==1), aes(y = avg_eff)) +
scale_y_log10(labels = scales::label_percent(accuracy = 0.01)) +
scale_x_log10(breaks = c(100, 1000, 10000, 50000), labels = c('100 bp', '1 kb', '10 kb', '50 kb'))+
labs(y = 'Efficiency', x = 'Deletion Size')+
scale_color_viridis_d()
plot_gal_aa_sizes
Fig. 4C - Deletion size accuracy
Again galK was done separately from the other loci, so let’s read in that data first:
df_galK_pin <- read_csv("../../data/low_throughput_experiments/2022_11_09_galK_sizes_3_accuracy.csv") %>%
mutate(accuracy = 1-(selective_colonies / permissive_colonies)) %>%
group_by(locus, deletion_size) %>%
mutate(avg_accuracy = mean(accuracy))
df_galK_pin## # A tibble: 12 × 9
## # Groups: locus, deletion_size [6]
## condition locus deletion_size pInt_…¹ repli…² permi…³ selec…⁴ accur…⁵ avg_a…⁶
## <chr> <chr> <chr> <lgl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 galK_1kb galK 1122 FALSE 1 8 0 1 1
## 2 galK_1kb galK 1122 FALSE 2 8 0 1 1
## 3 galK_4kb galK 4258 FALSE 1 8 0 1 0.938
## 4 galK_4kb galK 4258 FALSE 2 8 1 0.875 0.938
## 5 galK_10kb galK 10762 FALSE 1 8 1 0.875 0.875
## 6 galK_10kb galK 10762 FALSE 2 8 1 0.875 0.875
## 7 galK_24kb galK 24027 FALSE 1 8 0 1 1
## 8 galK_24kb galK 24027 FALSE 2 8 0 1 1
## 9 galK_49kb galK 49068 FALSE 1 8 1 0.875 0.875
## 10 galK_49kb galK 49068 FALSE 2 8 1 0.875 0.875
## 11 pInt_only galK - TRUE 1 8 8 0 0.0625
## 12 pInt_only galK - TRUE 2 8 7 0.125 0.0625
## # … with abbreviated variable names ¹pInt_only, ²replicate,
## # ³permissive_colonies, ⁴selective_colonies, ⁵accuracy, ⁶avg_accuracyThen we will read in the data for hisA, metA and leuD:
df_aa_pin <- read_csv("../../data/low_throughput_experiments/2022_09_07_AA_sizes_pin_plate_count.csv") %>%
mutate(accuracy = 1-(selective_colonies / permissive_colonies)) %>%
group_by(locus, deletion_size) %>%
mutate(avg_accuracy = mean(accuracy)) %>%
mutate(locus = factor(locus, levels = c('hisA','metA','leuD')))
df_aa_pin## # A tibble: 18 × 9
## # Groups: locus, deletion_size [9]
## condition locus delet…¹ pInt_…² repli…³ permi…⁴ selec…⁵ accur…⁶ avg_a…⁷
## <chr> <fct> <dbl> <lgl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 hisA 100 bp del hisA 100 FALSE 1 7 0 1 1
## 2 hisA 100 bp del hisA 100 FALSE 2 7 0 1 1
## 3 hisA 700 bp del hisA 700 FALSE 1 7 0 1 1
## 4 hisA 700 bp del hisA 700 FALSE 2 7 0 1 1
## 5 hisA 7000 bp d… hisA 7000 FALSE 1 7 0 1 0.929
## 6 hisA 7000 bp d… hisA 7000 FALSE 2 7 1 0.857 0.929
## 7 metA 100 bp del metA 100 FALSE 1 6 0 1 1
## 8 metA 100 bp del metA 100 FALSE 2 6 0 1 1
## 9 metA 900 bp del metA 900 FALSE 1 7 0 1 1
## 10 metA 900 bp del metA 900 FALSE 2 6 0 1 1
## 11 metA 13000 bp … metA 13000 FALSE 1 7 0 1 1
## 12 metA 13000 bp … metA 13000 FALSE 2 7 0 1 1
## 13 leuD 100 bp del leuD 100 FALSE 1 7 0 1 1
## 14 leuD 100 bp del leuD 100 FALSE 2 6 0 1 1
## 15 leuD 600 bp del leuD 600 FALSE 1 7 0 1 0.929
## 16 leuD 600 bp del leuD 600 FALSE 2 7 1 0.857 0.929
## 17 leuD 6000 bp d… leuD 6000 FALSE 1 7 0 1 0.929
## 18 leuD 6000 bp d… leuD 6000 FALSE 2 7 1 0.857 0.929
## # … with abbreviated variable names ¹deletion_size, ²pInt_only, ³replicate,
## # ⁴permissive_colonies, ⁵selective_colonies, ⁶accuracy, ⁷avg_accuracyWe will combine the data and plot:
#Combine data
df_galK_aa_pin <- bind_rows(df_galK_pin %>% filter(condition != 'pInt_only') %>% mutate(deletion_size = as.numeric(deletion_size)), df_aa_pin)
#Get background values
galK_bg <- (df_galK_pin %>% filter(condition == 'pInt_only' & replicate == 1))$avg_accuracy
#Plot with individual datapoints, and mean points connected by lines
plot_del_accuracy <- ggplot(df_galK_aa_pin, aes(x = deletion_size, y = accuracy, color = locus)) +
geom_hline(yintercept = 0, linetype = 2, color = 'light gray')+ #background accuracy for AA loci was 0.
geom_hline(yintercept = galK_bg, linetype = 4, color = 'light gray')+ #background accuracy for galK is shown separately here.
geom_line(data = . %>% filter(replicate==1), aes(y = avg_accuracy)) +
geom_jitter(shape = 21, width = 0.05, height = 0, alpha = 0.4) +
geom_point(data = . %>% filter(replicate==1), aes(y = avg_accuracy)) +
scale_y_continuous(labels = scales::label_percent(accuracy = 1), limits = c(0,1)) +
scale_x_log10(breaks = c(100, 1000, 10000, 50000), labels = c('100 bp', '1 kb', '10 kb', '50 kb'))+
labs(y = 'Phenotypic accuracy', x = 'Deletion Size')+
scale_color_viridis_d()
plot_del_accuracy
Fig. 4E - Integrating plasmid size efficiency (violacein operon)
For this experiment different sized fragments from the luxR inducible violacein operon were cloned into an integrating plasmid. Efficiency was measured for each plasmid with the ∆galK TO. Let’s read in the data:
df_vio <- read_csv("../../data/low_throughput_experiments/2022_10_25_pInt_vio_sizes.csv") %>%
mutate(eff = Kan_count / LB_count) %>%
group_by(condition, to, plasmid_size) %>% mutate(avg_eff = mean(eff)) #calculate efficiency and average efficiency for replicates
df_vio %>% kable() %>% kable_styling() %>% scroll_box(height = '250px')| condition | to | plasmid_size | replicate | LB_count | Kan_count | eff | avg_eff |
|---|---|---|---|---|---|---|---|
| p265 + pInt_kan | 265 | 1958 | 1 | 1140000 | 490 | 0.0004298 | 0.0004158 |
| p265 + pInt_kan | 265 | 1958 | 2 | 1120000 | 450 | 0.0004018 | 0.0004158 |
| p265 + pInt_vioA | 265 | 3628 | 1 | 1080000 | 590 | 0.0005463 | 0.0005694 |
| p265 + pInt_vioA | 265 | 3628 | 2 | 1080000 | 640 | 0.0005926 | 0.0005694 |
| p265 + pInt_vioAB | 265 | 6718 | 1 | 1120000 | 360 | 0.0003214 | 0.0002468 |
| p265 + pInt_vioAB | 265 | 6718 | 2 | 1220000 | 210 | 0.0001721 | 0.0002468 |
| p265 + pInt_vioABC | 265 | 8203 | 1 | 1210000 | 53 | 0.0000438 | 0.0000418 |
| p265 + pInt_vioABC | 265 | 8203 | 2 | 1280000 | 51 | 0.0000398 | 0.0000418 |
| pInt_vioABC only | control | 8203 | 1 | 1040000 | 2 | 0.0000019 | 0.0000019 |
| p265 + pInt_luxR_vioAE | 265 | 10671 | 1 | 1290000 | 32 | 0.0000248 | 0.0000230 |
| p265 + pInt_luxR_vioAE | 265 | 10671 | 2 | 1130000 | 24 | 0.0000212 | 0.0000230 |
| pInt_luxR_vioAE only | control | 10671 | 1 | 1400000 | 0 | 0.0000000 | 0.0000000 |
Now let’s plot:
plot_vio <- ggplot(df_vio%>% filter(to != 'control'), aes(x = plasmid_size, y = eff)) +
geom_jitter(shape = 21, width = 200, height = 0, color = "#440154FF", alpha = 0.4) +
geom_jitter(data = df_vio %>% filter(to == 'control'), shape = 4, width = 0, height = 0, color = 'light gray') + #negative control efficiencies for pInt_vioA-C and pInt_luxR_vioA-E are shown here
geom_point(data = . %>% filter(replicate ==1), aes(y = avg_eff), color = "#440154FF") +
geom_line(data = . %>% filter(replicate ==1), aes(y = avg_eff, group = to), color = "#440154FF") +
scale_y_log10(limits = c(NA,0.001), labels = scales::label_percent(accuracy = 0.001))+
scale_x_continuous(breaks = c(2000, 6000, 10000), labels = c('2 kb', '6 kb', '10 kb'))+
labs(x = 'Integrating plasmid size', y = 'Efficiency')
plot_vio
Create Fig. 4
theme_set(theme_figure())
fig_4_sizes <- plot_grid(plot_gal_aa_sizes + guides(color = 'none'), plot_vio, nrow = 1,
align = 'hv', axis = 'lr', scale = 1,
labels = c('B','E'))
fig_4_sizes <- plot_grid(plot_gal_aa_sizes + guides(color = 'none'),plot_del_accuracy + guides(color = 'none'),
plot_vio,NULL,
nrow = 2, ncol = 2,
align = 'hv', axis = 'lr', scale = 1,
labels = c('B','C','E'))
fig_4_sizes
save_plot("../../figures/r_pdf_figs/main_figs/fig_4_del_int_sizes.pdf", fig_4_sizes, base_width = 5, base_height = 4)sessionInfo()## R version 4.2.0 (2022-04-22)
## Platform: x86_64-apple-darwin17.0 (64-bit)
## Running under: macOS Big Sur/Monterey 10.16
##
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib
##
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] kableExtra_1.3.4 cowplot_1.1.1 viridis_0.6.2 viridisLite_0.4.1
## [5] knitr_1.41 forcats_0.5.2 stringr_1.5.0 dplyr_1.1.0
## [9] purrr_0.3.5 readr_2.1.3 tidyr_1.2.1 tibble_3.1.8
## [13] ggplot2_3.4.0 tidyverse_1.3.2
##
## loaded via a namespace (and not attached):
## [1] httr_1.4.4 sass_0.4.4 bit64_4.0.5
## [4] vroom_1.6.0 jsonlite_1.8.3 modelr_0.1.10
## [7] bslib_0.4.1 assertthat_0.2.1 highr_0.9
## [10] googlesheets4_1.0.1 cellranger_1.1.0 yaml_2.3.6
## [13] pillar_1.8.1 backports_1.4.1 glue_1.6.2
## [16] digest_0.6.30 rvest_1.0.3 colorspace_2.0-3
## [19] htmltools_0.5.4 pkgconfig_2.0.3 broom_1.0.1
## [22] haven_2.5.1 scales_1.2.1 webshot_0.5.4
## [25] svglite_2.1.0 tzdb_0.3.0 timechange_0.1.1
## [28] googledrive_2.0.0 generics_0.1.3 farver_2.1.1
## [31] ellipsis_0.3.2 cachem_1.0.6 withr_2.5.0
## [34] cli_3.4.1 magrittr_2.0.3 crayon_1.5.2
## [37] readxl_1.4.1 evaluate_0.18 fs_1.5.2
## [40] fansi_1.0.3 xml2_1.3.3 textshaping_0.3.6
## [43] tools_4.2.0 hms_1.1.2 gargle_1.2.1
## [46] lifecycle_1.0.3 munsell_0.5.0 reprex_2.0.2
## [49] compiler_4.2.0 jquerylib_0.1.4 systemfonts_1.0.4
## [52] rlang_1.0.6 grid_4.2.0 rstudioapi_0.14
## [55] labeling_0.4.2 rmarkdown_2.18 gtable_0.3.1
## [58] DBI_1.1.3 R6_2.5.1 gridExtra_2.3
## [61] lubridate_1.9.0 fastmap_1.1.0 bit_4.0.5
## [64] utf8_1.2.2 ragg_1.2.5 stringi_1.7.8
## [67] parallel_4.2.0 vctrs_0.5.2 dbplyr_2.2.1
## [70] tidyselect_1.2.0 xfun_0.35