library(tidyverse)
library(cowplot)
library(broom) 
library(modelr) 
library(viridis)
library(lubridate)
library(hms)
library(knitr)
library(kableExtra)

knitr::opts_chunk$set(tidy.opts=list(width.cutoff=60),tidy=TRUE, echo = TRUE, message=FALSE, warning=FALSE, fig.align="center")

source("../../tools/echem_processing_tools.R")
source("../../tools/plotting_tools.R")

theme_set(theme_1())

Data

# Agarose
gc_agarose <- read_csv("../../06_06_19_agarose_PYO_2/Processing/06_06_19_processed_gc_max_agarose.csv")
swv_agarose <- read_csv("../../06_06_19_agarose_PYO_2/Processing/06_06_19_processed_swv_max_agarose.csv")
swvGC_agarose <- read_csv("../../06_06_19_agarose_PYO_2/Processing/06_06_19_processed_swvGC_agarose.csv")

# Agarose + ctDNA

gc_ctDNA <- read_csv("../processing/08_13_19_processed_gc_max_ctDNA.csv")
swv_ctDNA <- read_csv("../processing/08_13_19_processed_swv_max_ctDNA.csv")
swvGC_ctDNA <- read_csv("../processing/08_13_19_processed_swvGC_ctDNA.csv")

GC time course

Overview

df_gc <- bind_rows(gc_agarose %>% mutate(condition = "agarose"), 
    gc_ctDNA %>% mutate(condition = "ctDNA")) %>% group_by(condition, 
    reactor) %>% mutate(min_time = min(minutes)) %>% mutate(norm_time = minutes - 
    min_time)

ggplot(df_gc, aes(x = norm_time, y = max_current, color = reactor, 
    shape = electrode)) + geom_point() + facet_wrap(~condition)

Transfer runs

ggplot(df_gc %>% filter(norm_time < 100) %>% filter(reactor != 
    "soak" & reactor != "soak_2"), aes(x = norm_time, y = max_current, 
    color = condition, shape = electrode)) + geom_point() + facet_wrap(~electrode)

Looks like their may actually be a small difference especially when normalizing by the starting values…

Soak Runs

ggplot(df_gc %>% filter(norm_time < 100) %>% filter(reactor == 
    "soak" | reactor == "soak_2"), aes(x = norm_time, y = max_current, 
    color = condition, shape = electrode)) + geom_point() + facet_wrap(~electrode)

Hard to tell without normalization / letting it go longer…but ctDNA may actually be slower.

SWV time course

Overview

df_swv <- bind_rows(swv_agarose %>% mutate(condition = "agarose"), 
    swv_ctDNA %>% mutate(condition = "ctDNA")) %>% group_by(condition, 
    reactor) %>% mutate(min_time = min(minutes)) %>% mutate(norm_time = minutes - 
    min_time)

ggplot(df_swv, aes(x = norm_time, y = max_current, color = reactor, 
    shape = electrode)) + geom_point() + facet_wrap(~condition)

Transfer Runs

ggplot(df_swv %>% filter(reactor != "soak" & reactor != "soak_2") %>% 
    filter(norm_time < 100), aes(x = norm_time, y = max_current, 
    color = reactor, shape = electrode)) + geom_point() + facet_wrap(~electrode)

Again, looks like the ctDNA may actually be slower, especially after normalizing. Transfer 2 / 3 are replicates and look pretty similar, so that’s good.

Soak Runs

ggplot(df_swv %>% filter(reactor == "soak" | reactor == "soak_2"), 
    aes(x = norm_time, y = max_current, color = reactor, shape = electrode)) + 
    geom_point() + facet_wrap(~electrode)

Same as before…a little hard to tell without normalization or running longer. ctDNA is possibly slower.

SWV vs. GC

Overview

df_swv_gc <- bind_rows(swvGC_agarose %>% mutate(condition = "agarose"), 
    swvGC_ctDNA %>% mutate(condition = "ctDNA"))

ggplot(df_swv_gc, aes(x = max_current_swv, y = max_current_gc, 
    color = reactor)) + geom_point() + facet_wrap(~electrode_swv) + 
    scale_color_viridis_d() + geom_smooth(data = df_swv_gc %>% 
    filter(rep <= 10), method = "lm")

Overall slopes look pretty similar. Transfer_2 is an obvious outlier with a shallower slope.

Conclusions