Violation of the pairwise assembly rule in 3 and 4-member communities

Author

Shane Hogle

Published

July 4, 2025

Abstract

This notebook examines the experiments from 3- and 4-member sets that violate the pairwise assembly rule.

The pairwise asembly rule states “that in a multispecies competition, species that allcoexist with each other in pairs will survive, whereas species that are excluded by any of the surviving species will go extinct” (Friedman, 2017). Here we simply collect the competition outcomes from the pairwise experiments into a binary matrix, then we decompose the coexisting members (species with f > 1% | f < 99% after 8x 48 hour transfers) from each 3 or 4-member set into all unique constituent pairs, then we finally check whether those pairs coexist in in the pairwise binary matrix.

1 Setup

1.1 Libraries

Show/hide code

library(tidyverse)
library(here)
library(fs)
library(scales)
source(here::here("R", "communities", "community_composition", "utils_community_composition.R"))

1.2 Global variables

Show/hide code

data <- here::here("data", "communities")
# make processed data directory if it doesn't exist
fs::dir_create(data)

2 Read data

2.1 Pairwise coexistence outcomes

Show/hide code

outcome_pairs <- read_tsv(here::here(data, "2sp_coexistence.tsv"))

Rows: 95 Columns: 4
── Column specification ────────────────────────────────────────────────────────
Delimiter: "\t"
chr (2): sp_1, sp_2
dbl (2): strep_conc, coexists

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Convert the pairs in tibble format into the matrixes

Show/hide code

pairs_matrixified <- outcome_pairs %>%
  nest(data = -strep_conc) %>% 
  mutate(mat = map(data, matrixify))

2.2 Species abundances

Show/hide code

samp_trios <- readr::read_tsv(here::here(data, "3sps_compiled.tsv")) %>% 
  dplyr::rename(f = f_thresh) %>% 
  dplyr::filter(community_type == "experiment") %>% 
  dplyr::mutate(sp = paste(str_to_upper(evo_hist), str_extract(strainID, "\\d+"), sep = "_")) %>% 
  dplyr::select(sample, strep_conc, replicate, n_species, sp, f, start_f=target_f_masterplate)

samp_quarts <- readr::read_tsv(here::here(data, "4sps_compiled.tsv")) %>% 
  dplyr::rename(f = f_thresh) %>% 
  dplyr::filter(community_type == "experiment") %>% 
  dplyr::mutate(sp = paste(str_to_upper(evo_hist), str_extract(strainID, "\\d+"), sep = "_"))

samp_octs <- readr::read_tsv(here::here(data, "8sps_compiled.tsv")) %>% 
  dplyr::rename(f = f_thresh) %>% 
  dplyr::filter(community_type == "experiment") %>% 
  mutate(target_f_masterplate = target_f_masterplate*100, 
         max_f = max_f*100)

3 Trios: Application of the pairwise assembly rule

For species trios there are 32 different strain combinations. 4 all ancestral trios, 4 all evolved trios, and 24 trios containing a mixture of evolved and ancestral species

Show/hide code

samp_trios_fmt <- samp_trios %>% 
  group_by(sample, strep_conc, replicate, n_species) %>% 
  # for each sample create two new vars that record the species that went into
  # the trio and their rough relative abundance
  mutate(sp_set_id = paste(sp, collapse = ", "),
         sp_start_f = paste(start_f, collapse = ", ")) %>% 
  # The next filter step ensures we remove species that went extinct and species
  # that were the only species left in the community (e.g., samples where 1287
  # was 100%) now we only look at species coexisting in each sample where there
  # are at least two species coexisting
  filter(f >= 0.01 & f <= 0.99) %>% 
  # record as a variable a comma separated character of the rounded final
  # species abundances
  mutate(sp_final_f = paste(round(f, 2), collapse = ", ")) %>% 
  # record a character variable of comma separated species that were found
  # coexisting at the final time point
  mutate(coexisting = paste(sp, collapse = ", ")) %>% 
  ungroup() %>% 
  dplyr::select(sample, strep_conc, replicate, n_species, sp_set_id, sp_start_f, sp_final_f, coexisting) %>% 
  # removes redundant rows
  distinct() %>% 
  # now use nest and strsplit so that we can record a character vector of coexisting species 
  nest(.by = c(sample, strep_conc, replicate)) %>% 
  mutate(coexisting = map(data, ~ strsplit(.x$coexisting, split=", ", fixed=TRUE)[[1]])) %>% 
  #mutate(sp_set_id = map(data, ~strsplit(.x$sp_set_id, split=", ", fixed=TRUE)[[1]])) %>% 
  # join with the pairwise coexistence matrixes
  left_join(pairs_matrixified, by = join_by(strep_conc))

Apply the coexistence rule test

Show/hide code

samp_trios_ruletest <- samp_trios_fmt %>% 
  mutate(ruletest = map2(coexisting, mat, coexistence_rule))

samp_trios_ruletest_reduced <- samp_trios_ruletest %>% 
  select(sample, strep_conc, replicate, data.x, ruletest) %>% 
  unnest(cols = c(data.x, ruletest)) %>% 
  dplyr::select(-coexisting)

Show/hide code

distinct(samp_trios, sample) %>% count()

There are 922 trio samples total

Show/hide code

samp_trios_ruletest_reduced %>% 
  filter(pair_coexists_alone == 0) %>% 
  distinct(sample) %>% 
  count()

291 Trio samples (291/922 = 31%) have at least 1 species pair combination that is inconsistent with outcomes from pairwise competition

We can compare the distribution of these violations across different antibiotics levels and evolutionary histories

Show/hide code

# number of discrete samples with varying species starting frequencies for 
# each species/strain combination
nreps <- 12

full_set <- samp_trios_ruletest_reduced %>% 
  distinct(strep_conc, sp_set_id) %>% 
  complete(sp_set_id, strep_conc)

samp_trios_ruletest_reduced_summary <- samp_trios_ruletest_reduced %>% 
  filter(pair_coexists_alone == 0) %>% 
  distinct(sample, strep_conc, replicate, n_species, sp_set_id) %>%
  group_by(strep_conc, sp_set_id) %>% 
  count(name = "n_trios_w_pair_violation") %>% 
  right_join(full_set, by = join_by(strep_conc, sp_set_id)) %>% 
  replace_na(list(n_trios_w_pair_violation = 0)) %>% 
  ungroup() %>% 
  mutate(perc = n_trios_w_pair_violation/nreps*100, 
         strep_conc = factor(strep_conc)) %>% 
  mutate(evo_combo = case_when(str_detect(sp_set_id, "ANC") & str_detect(sp_set_id, "EVO") ~ "mix",
                               str_detect(sp_set_id, "ANC") &! str_detect(sp_set_id, "EVO") ~ "all_anc",
                               str_detect(sp_set_id, "EVO") &! str_detect(sp_set_id, "ANC") ~ "all_evo"))

Show/hide code

samp_trios_ruletest_reduced_summary %>% 
  filter(evo_combo == "all_anc") %>% 
  plot_heatmap(label = n_trios_w_pair_violation) +
  labs(title = "3-member sets, uniform histories\n(ANC/strep sensitive)")

Show/hide code

samp_trios_ruletest_reduced_summary %>% 
  filter(evo_combo == "all_evo") %>% 
  plot_heatmap(label = n_trios_w_pair_violation) +
  labs(title = "3-member sets, uniform histories\n(EVO/strep resistant)")

Show/hide code

samp_trios_ruletest_reduced_summary %>% 
  filter(evo_combo == "mix") %>% 
  plot_heatmap(label = n_trios_w_pair_violation) +
  labs(title = "3-member sets, mixed histories\n(ANC+EVO/strep sensitive + resistant)")

Quick check into relative fraction of each grouping (all sensitive, all resistant, mixed sensitive/resistant) with at least one violation to the asse

Show/hide code

a <- samp_trios_ruletest_reduced_summary %>% 
  group_by(strep_conc, evo_combo) %>% 
  count(name = "ntot")

b <- samp_trios_ruletest_reduced_summary %>% 
  group_by(strep_conc, evo_combo) %>% 
  count(n_trios_w_pair_violation != 0)

left_join(b, a, by = join_by(strep_conc, evo_combo)) %>% 
  mutate(f = n/ntot) %>% 
  ggplot() +
  geom_col(aes(x = strep_conc, y = f, fill = `n_trios_w_pair_violation != 0`)) +
  facet_grid(~ evo_combo) +
  scale_x_discrete(guide = guide_axis(angle = 0)) +
  scale_y_continuous(labels = label_percent()) +
  labs(x = "Streptomycin concentration (ug/ml)", y = "Percentage of trio sets",
       fill = "At least one experiment\nviolating assembly rule")

4 Quartets: Application of the pairwise assembly rule

For species quartets there are 16 different strain combinations. 1 all ancestral, 1 all evolved, and 14 containing a mixture of evolved and ancestral

Show/hide code

samp_quarts_fmt <- samp_quarts %>% 
  group_by(sample, strep_conc, replicate, n_species) %>% 
  # for each sample create two new vars that record the species that went into
  # the trio and their rough relative abundance
  mutate(sp_set_id = paste(sp, collapse = ", "),
         sp_start_f = paste(target_f_masterplate, collapse = ", ")) %>% 
  # The next filter step ensures we remove species that went extinct and species
  # that were the only species left in the community (e.g., samples where 1287
  # was 100%) now we only look at species coexisting in each sample where there
  # are at least two species coexisting
  filter(f >= 0.01 & f <= 0.99) %>% 
  # record as a variable a comma separated character of the rounded final
  # species abundances
  mutate(sp_final_f = paste(round(f, 2), collapse = ", ")) %>% 
  # record a character variable of comma separated species that were found
  # coexisting at the final time point
  mutate(coexisting = paste(sp, collapse = ", ")) %>% 
  ungroup() %>% 
  dplyr::select(sample, strep_conc, replicate, n_species, sp_set_id, sp_start_f, sp_final_f, coexisting) %>% 
  # removes redundant rows
  distinct() %>% 
  # now use nest and strsplit so that we can record a character vector of coexisting species 
  nest(.by = c(sample, strep_conc, replicate)) %>% 
  mutate(coexisting = map(data, ~ strsplit(.x$coexisting, split=", ", fixed=TRUE)[[1]])) %>% 
  # join with the pairwise coexistence matrixes
  left_join(pairs_matrixified, by = join_by(strep_conc))

Apply the coexistence rule test

Show/hide code

samp_quarts_ruletest <- samp_quarts_fmt %>% 
  mutate(ruletest = map2(coexisting, mat, coexistence_rule))

samp_quarts_ruletest_reduced <- samp_quarts_ruletest %>% 
  select(sample, strep_conc, replicate, data.x, ruletest) %>% 
  unnest(cols = c(data.x, ruletest)) %>% 
  dplyr::select(-coexisting)

Show/hide code

distinct(samp_quarts, sample) %>% count()

There are 512 quartet samples total

Show/hide code

samp_quarts_ruletest_reduced %>% 
  filter(pair_coexists_alone == 0) %>% 
  distinct(sample) %>% 
  count()

116 Trio samples (116/512 = 23%) have at least 1 species pair combination that is inconsistent with outcomes from pairwise competition

We can compare the distribution of these violations across different antibiotics levels and evolutionary histories

Show/hide code

# number of discrete samples with varying species starting frequencies for 
# each species/strain combination
nreps <- 8

full_set <- samp_quarts_ruletest_reduced %>% 
  distinct(strep_conc, sp_set_id) %>% 
  complete(sp_set_id, strep_conc)

samp_quarts_ruletest_reduced_summary <- samp_quarts_ruletest_reduced %>% 
  filter(pair_coexists_alone == 0) %>% 
  distinct(sample, strep_conc, replicate, n_species, sp_set_id) %>%
  group_by(strep_conc, sp_set_id) %>% 
  count(name = "n_quarts_w_pair_violation") %>% 
  right_join(full_set, by = join_by(strep_conc, sp_set_id)) %>% 
  replace_na(list(n_quarts_w_pair_violation = 0)) %>% 
  ungroup() %>% 
  mutate(perc = round(n_quarts_w_pair_violation/nreps*100), 
         strep_conc = factor(strep_conc)) %>% 
  mutate(evo_combo = case_when(str_detect(sp_set_id, "ANC") & str_detect(sp_set_id, "EVO") ~ "mix",
                               str_detect(sp_set_id, "ANC") &! str_detect(sp_set_id, "EVO") ~ "all_anc",
                               str_detect(sp_set_id, "EVO") &! str_detect(sp_set_id, "ANC") ~ "all_evo"))

Show/hide code

samp_quarts_ruletest_reduced_summary %>% 
  filter(evo_combo != "mix") %>% 
  plot_heatmap(label = n_quarts_w_pair_violation) +
  labs(title = "4-member sets, uniform histories")

Show/hide code

samp_quarts_ruletest_reduced_summary %>% 
  filter(evo_combo == "mix") %>% 
  plot_heatmap(label = n_quarts_w_pair_violation) +
  labs(title = "4-member sets, mixed histories")

Quick check into relative fraction of each grouping (all sensitive, all resistant, mixed sensitive/resistant) with at least one violation to the asse

Show/hide code

a <- samp_quarts_ruletest_reduced_summary %>% 
  group_by(strep_conc, evo_combo) %>% 
  count(name = "ntot")

b <- samp_quarts_ruletest_reduced_summary %>% 
  group_by(strep_conc, evo_combo) %>% 
  count(n_quarts_w_pair_violation != 0)

left_join(b, a, by = join_by(strep_conc, evo_combo)) %>% 
  mutate(f = n/ntot) %>% 
  ggplot() +
  geom_col(aes(x = strep_conc, y = f, fill = `n_quarts_w_pair_violation != 0`)) +
  facet_grid(~ evo_combo) +
  scale_x_discrete(guide = guide_axis(angle = 0)) +
  scale_y_continuous(labels = label_percent()) +
  labs(x = "Streptomycin concentration (ug/ml)", y = "Percentage of 4-member sets",
       fill = "At least one experiment\nviolating assembly rule")

--- title: "Violation of the pairwise assembly rule in 3 and 4-member communities" author: "Shane Hogle" date: today link-citations: true abstract: "This notebook examines the experiments from 3- and 4-member sets that violate the pairwise assembly rule." --- The pairwise asembly rule states "that in a multispecies competition, species that allcoexist with each other in pairs will survive, whereas species that are excluded by any of the surviving species will go extinct" (Friedman, 2017). Here we simply collect the competition outcomes from the pairwise experiments into a binary matrix, then we decompose the coexisting members (species with f > 1% | f < 99% after 8x 48 hour transfers) from each 3 or 4-member set into all unique constituent pairs, then we finally check whether those pairs coexist in in the pairwise binary matrix. # Setup ## Libraries ```{r} #| output: false library(tidyverse) library(here) library(fs) library(scales) source(here::here("R", "communities", "community_composition", "utils_community_composition.R")) ``` ## Global variables ```{r} data <- here::here("data", "communities") # make processed data directory if it doesn't exist fs::dir_create(data) ``` # Read data ## Pairwise coexistence outcomes ```{r} outcome_pairs <- read_tsv(here::here(data, "2sp_coexistence.tsv")) ``` Convert the pairs in tibble format into the matrixes ```{r} pairs_matrixified <- outcome_pairs %>% nest(data = -strep_conc) %>% mutate(mat = map(data, matrixify)) ``` ## Species abundances ```{r} #| output: false samp_trios <- readr::read_tsv(here::here(data, "3sps_compiled.tsv")) %>% dplyr::rename(f = f_thresh) %>% dplyr::filter(community_type == "experiment") %>% dplyr::mutate(sp = paste(str_to_upper(evo_hist), str_extract(strainID, "\\d+"), sep = "_")) %>% dplyr::select(sample, strep_conc, replicate, n_species, sp, f, start_f=target_f_masterplate) samp_quarts <- readr::read_tsv(here::here(data, "4sps_compiled.tsv")) %>% dplyr::rename(f = f_thresh) %>% dplyr::filter(community_type == "experiment") %>% dplyr::mutate(sp = paste(str_to_upper(evo_hist), str_extract(strainID, "\\d+"), sep = "_")) samp_octs <- readr::read_tsv(here::here(data, "8sps_compiled.tsv")) %>% dplyr::rename(f = f_thresh) %>% dplyr::filter(community_type == "experiment") %>% mutate(target_f_masterplate = target_f_masterplate*100, max_f = max_f*100) ``` # Trios: Application of the pairwise assembly rule For species trios there are 32 different strain combinations. 4 all ancestral trios, 4 all evolved trios, and 24 trios containing a mixture of evolved and ancestral species ```{r} samp_trios_fmt <- samp_trios %>% group_by(sample, strep_conc, replicate, n_species) %>% # for each sample create two new vars that record the species that went into # the trio and their rough relative abundance mutate(sp_set_id = paste(sp, collapse = ", "), sp_start_f = paste(start_f, collapse = ", ")) %>% # The next filter step ensures we remove species that went extinct and species # that were the only species left in the community (e.g., samples where 1287 # was 100%) now we only look at species coexisting in each sample where there # are at least two species coexisting filter(f >= 0.01 & f <= 0.99) %>% # record as a variable a comma separated character of the rounded final # species abundances mutate(sp_final_f = paste(round(f, 2), collapse = ", ")) %>% # record a character variable of comma separated species that were found # coexisting at the final time point mutate(coexisting = paste(sp, collapse = ", ")) %>% ungroup() %>% dplyr::select(sample, strep_conc, replicate, n_species, sp_set_id, sp_start_f, sp_final_f, coexisting) %>% # removes redundant rows distinct() %>% # now use nest and strsplit so that we can record a character vector of coexisting species nest(.by = c(sample, strep_conc, replicate)) %>% mutate(coexisting = map(data, ~ strsplit(.x$coexisting, split=", ", fixed=TRUE)[[1]])) %>% #mutate(sp_set_id = map(data, ~strsplit(.x$sp_set_id, split=", ", fixed=TRUE)[[1]])) %>% # join with the pairwise coexistence matrixes left_join(pairs_matrixified, by = join_by(strep_conc)) ``` Apply the coexistence rule test ```{r} samp_trios_ruletest <- samp_trios_fmt %>% mutate(ruletest = map2(coexisting, mat, coexistence_rule)) samp_trios_ruletest_reduced <- samp_trios_ruletest %>% select(sample, strep_conc, replicate, data.x, ruletest) %>% unnest(cols = c(data.x, ruletest)) %>% dplyr::select(-coexisting) ``` ```{r} distinct(samp_trios, sample) %>% count() ``` There are 922 trio samples total ```{r} samp_trios_ruletest_reduced %>% filter(pair_coexists_alone == 0) %>% distinct(sample) %>% count() ``` 291 Trio samples (291/922 = 31%) have at least 1 species pair combination that is inconsistent with outcomes from pairwise competition We can compare the distribution of these violations across different antibiotics levels and evolutionary histories ```{r} # number of discrete samples with varying species starting frequencies for # each species/strain combination nreps <- 12 full_set <- samp_trios_ruletest_reduced %>% distinct(strep_conc, sp_set_id) %>% complete(sp_set_id, strep_conc) samp_trios_ruletest_reduced_summary <- samp_trios_ruletest_reduced %>% filter(pair_coexists_alone == 0) %>% distinct(sample, strep_conc, replicate, n_species, sp_set_id) %>% group_by(strep_conc, sp_set_id) %>% count(name = "n_trios_w_pair_violation") %>% right_join(full_set, by = join_by(strep_conc, sp_set_id)) %>% replace_na(list(n_trios_w_pair_violation = 0)) %>% ungroup() %>% mutate(perc = n_trios_w_pair_violation/nreps*100, strep_conc = factor(strep_conc)) %>% mutate(evo_combo = case_when(str_detect(sp_set_id, "ANC") & str_detect(sp_set_id, "EVO") ~ "mix", str_detect(sp_set_id, "ANC") &! str_detect(sp_set_id, "EVO") ~ "all_anc", str_detect(sp_set_id, "EVO") &! str_detect(sp_set_id, "ANC") ~ "all_evo")) ``` ::: {#fig-01} ```{r} #| label: fig-01 #| fig-width: 7 #| fig-height: 5 #| warning: false samp_trios_ruletest_reduced_summary %>% filter(evo_combo == "all_anc") %>% plot_heatmap(label = n_trios_w_pair_violation) + labs(title = "3-member sets, uniform histories\n(ANC/strep sensitive)") ``` The composition of each species/strain combination (y-axis) was assessed in four different streptomycin concentrations (x-axis, 0, 16, 64, 256 ug/ml) after eight 48 hour transfers in n=12 different microcosms with varying starting proportions of each species. The heatmap shows the percentage of the n=12 microcosms for each species/strain combination that violates the pairwise assembly rule while the absolute number of violations is shown in text. The species/strains trios here only contain the ancestral/streptomycin sensitive forms. ::: ::: {#fig-02} ```{r} #| label: fig-02 #| fig-width: 7 #| fig-height: 5 #| warning: false samp_trios_ruletest_reduced_summary %>% filter(evo_combo == "all_evo") %>% plot_heatmap(label = n_trios_w_pair_violation) + labs(title = "3-member sets, uniform histories\n(EVO/strep resistant)") ``` As in @fig-01 but the species/strains trios here only contain the evolved/streptomycin resistant forms. ::: ::: {#fig-03} ```{r} #| label: fig-03 #| fig-width: 5 #| fig-height: 8 #| warning: false samp_trios_ruletest_reduced_summary %>% filter(evo_combo == "mix") %>% plot_heatmap(label = n_trios_w_pair_violation) + labs(title = "3-member sets, mixed histories\n(ANC+EVO/strep sensitive + resistant)") ``` As in @fig-01 but the species/strains trios here only contain all possible mixtures of ancestra/evolved (sensitive/resistant) species. ::: Quick check into relative fraction of each grouping (all sensitive, all resistant, mixed sensitive/resistant) with at least one violation to the asse ::: {#fig-04} ```{r} #| label: fig-04 #| fig-width: 8 #| fig-height: 4 #| warning: false a <- samp_trios_ruletest_reduced_summary %>% group_by(strep_conc, evo_combo) %>% count(name = "ntot") b <- samp_trios_ruletest_reduced_summary %>% group_by(strep_conc, evo_combo) %>% count(n_trios_w_pair_violation != 0) left_join(b, a, by = join_by(strep_conc, evo_combo)) %>% mutate(f = n/ntot) %>% ggplot() + geom_col(aes(x = strep_conc, y = f, fill = `n_trios_w_pair_violation != 0`)) + facet_grid(~ evo_combo) + scale_x_discrete(guide = guide_axis(angle = 0)) + scale_y_continuous(labels = label_percent()) + labs(x = "Streptomycin concentration (ug/ml)", y = "Percentage of trio sets", fill = "At least one experiment\nviolating assembly rule") ``` Percentage of trio species sets containing at least one experiment violating the assembly rule. Horizontal axis shows the different streptomycin concentrations tested, while different panels show the all ancestral species trio sets (n=4), the all evolved species trio sets (n=4), and all possible mixtures of ancestral and evolved species trio sets (n=24). ::: # Quartets: Application of the pairwise assembly rule For species quartets there are 16 different strain combinations. 1 all ancestral, 1 all evolved, and 14 containing a mixture of evolved and ancestral ```{r} samp_quarts_fmt <- samp_quarts %>% group_by(sample, strep_conc, replicate, n_species) %>% # for each sample create two new vars that record the species that went into # the trio and their rough relative abundance mutate(sp_set_id = paste(sp, collapse = ", "), sp_start_f = paste(target_f_masterplate, collapse = ", ")) %>% # The next filter step ensures we remove species that went extinct and species # that were the only species left in the community (e.g., samples where 1287 # was 100%) now we only look at species coexisting in each sample where there # are at least two species coexisting filter(f >= 0.01 & f <= 0.99) %>% # record as a variable a comma separated character of the rounded final # species abundances mutate(sp_final_f = paste(round(f, 2), collapse = ", ")) %>% # record a character variable of comma separated species that were found # coexisting at the final time point mutate(coexisting = paste(sp, collapse = ", ")) %>% ungroup() %>% dplyr::select(sample, strep_conc, replicate, n_species, sp_set_id, sp_start_f, sp_final_f, coexisting) %>% # removes redundant rows distinct() %>% # now use nest and strsplit so that we can record a character vector of coexisting species nest(.by = c(sample, strep_conc, replicate)) %>% mutate(coexisting = map(data, ~ strsplit(.x$coexisting, split=", ", fixed=TRUE)[[1]])) %>% # join with the pairwise coexistence matrixes left_join(pairs_matrixified, by = join_by(strep_conc)) ``` Apply the coexistence rule test ```{r} samp_quarts_ruletest <- samp_quarts_fmt %>% mutate(ruletest = map2(coexisting, mat, coexistence_rule)) samp_quarts_ruletest_reduced <- samp_quarts_ruletest %>% select(sample, strep_conc, replicate, data.x, ruletest) %>% unnest(cols = c(data.x, ruletest)) %>% dplyr::select(-coexisting) ``` ```{r} distinct(samp_quarts, sample) %>% count() ``` There are 512 quartet samples total ```{r} samp_quarts_ruletest_reduced %>% filter(pair_coexists_alone == 0) %>% distinct(sample) %>% count() ``` 116 Trio samples (116/512 = 23%) have at least 1 species pair combination that is inconsistent with outcomes from pairwise competition We can compare the distribution of these violations across different antibiotics levels and evolutionary histories ```{r} # number of discrete samples with varying species starting frequencies for # each species/strain combination nreps <- 8 full_set <- samp_quarts_ruletest_reduced %>% distinct(strep_conc, sp_set_id) %>% complete(sp_set_id, strep_conc) samp_quarts_ruletest_reduced_summary <- samp_quarts_ruletest_reduced %>% filter(pair_coexists_alone == 0) %>% distinct(sample, strep_conc, replicate, n_species, sp_set_id) %>% group_by(strep_conc, sp_set_id) %>% count(name = "n_quarts_w_pair_violation") %>% right_join(full_set, by = join_by(strep_conc, sp_set_id)) %>% replace_na(list(n_quarts_w_pair_violation = 0)) %>% ungroup() %>% mutate(perc = round(n_quarts_w_pair_violation/nreps*100), strep_conc = factor(strep_conc)) %>% mutate(evo_combo = case_when(str_detect(sp_set_id, "ANC") & str_detect(sp_set_id, "EVO") ~ "mix", str_detect(sp_set_id, "ANC") &! str_detect(sp_set_id, "EVO") ~ "all_anc", str_detect(sp_set_id, "EVO") &! str_detect(sp_set_id, "ANC") ~ "all_evo")) ``` ::: {#fig-05} ```{r} #| label: fig-05 #| fig-width: 7 #| fig-height: 3 #| warning: false samp_quarts_ruletest_reduced_summary %>% filter(evo_combo != "mix") %>% plot_heatmap(label = n_quarts_w_pair_violation) + labs(title = "4-member sets, uniform histories") ``` The composition of each 4 member species/strain combination (y-axis) was assessed in four different streptomycin concentrations (x-axis, 0, 16, 64, 256 ug/ml) after eight 48 hour transfers in n=8 different microcosms with varying starting proportions of each species. The heatmap shows the percentage of the n=8 microcosms for each 4 member species/strain combination that violates the pairwise assembly rule while the absolute number of violations is shown in text. The species/strains quartets here only contain either all evolved/resistant or all ancestral/sensitive forms. ::: ::: {#fig-06} ```{r} #| label: fig-06 #| fig-width: 7 #| fig-height: 6 #| warning: false samp_quarts_ruletest_reduced_summary %>% filter(evo_combo == "mix") %>% plot_heatmap(label = n_quarts_w_pair_violation) + labs(title = "4-member sets, mixed histories") ``` As in @fig-05 but the 4-member sets here contain mixtures of ancestra/evolved (sensitive/resistant) species. ::: Quick check into relative fraction of each grouping (all sensitive, all resistant, mixed sensitive/resistant) with at least one violation to the asse ::: {#fig-07} ```{r} #| label: fig-07 #| fig-width: 8 #| fig-height: 4 #| warning: false a <- samp_quarts_ruletest_reduced_summary %>% group_by(strep_conc, evo_combo) %>% count(name = "ntot") b <- samp_quarts_ruletest_reduced_summary %>% group_by(strep_conc, evo_combo) %>% count(n_quarts_w_pair_violation != 0) left_join(b, a, by = join_by(strep_conc, evo_combo)) %>% mutate(f = n/ntot) %>% ggplot() + geom_col(aes(x = strep_conc, y = f, fill = `n_quarts_w_pair_violation != 0`)) + facet_grid(~ evo_combo) + scale_x_discrete(guide = guide_axis(angle = 0)) + scale_y_continuous(labels = label_percent()) + labs(x = "Streptomycin concentration (ug/ml)", y = "Percentage of 4-member sets", fill = "At least one experiment\nviolating assembly rule") ``` Percentage of 4 member sets containing at least one experiment violating the assembly rule. Horizontal axis shows the different streptomycin concentrations tested, while different panels show the all ancestral species trio sets (n=4), the all evolved species trio sets (n=4), and all possible mixtures of ancestral and evolved species trio sets (n=24). :::