Hello,
I've been working with the package version ‘3.12.0’ and have encountered some points of confusion regarding statistical results derived from bootstrap samples, specifically within the "synergy_score_statistics" and "response_statistics" dataframes.
To illustrate my concern, for the "synergy_score_statistics" dataframe, the confidence interval boundaries given by [Bliss_synergy_ci_left, Bliss_synergy_ci_right] do not coincide with what I expected from calculating [Bliss_synergy_mean - 1.96Bliss_synergy_sem, Bliss_synergy_mean + 1.96Bliss_synergy_sem]. This same pattern of discrepancy is observed with other algorithms' statistics as well.
Additionally, in the "response_statistics" dataframe produced by the CalculateSensitivity function, I noticed that response_mean seems to be the average of the replicated dose-response matrix rather than the mean of bootstrap samples. Correspondingly, the response_sd and response_sem are also seemingly not derived from bootstrap samples, which differs from the approach taken in the "synergy_score_statistics" dataframe. I am curious as to why bootstrap means are not used for response_statistics as they are for synergy_score_statistics. Moreover, the confidence interval boundaries [response_ci_left, response_ci_right] are incongruous with the calculated boundaries [response_mean ± 1.96*response_sem].
I'm still honing my statistical understanding and would greatly appreciate clarification on these matters.
Thank you very much for your time, and I look forward to your insights. Here is the code follow the instruction:
`data("ONEIL_screening_data")
res <- ReshapeData(
data = ONEIL_screening_data,
data_type = "inhibition",
impute = TRUE,
impute_method = NULL,
noise = TRUE,
iteration = 10, # Number of iterations for bootstrapping
seed = 1
)
synergy <- CalculateSynergy(
data = res,
method = c("ZIP", "HSA", "Bliss", "Loewe"),
Emin = NA,
Emax = NA,
adjusted = TRUE,
correct_baseline = "non",
iteration = 10,
seed = 123
)
sensitivity <- CalculateSensitivity(
data = res,
correct_baseline = "non",
iteration = 10 # Number of iterations for bootstrapping
)`
Hello,
I've been working with the package version ‘3.12.0’ and have encountered some points of confusion regarding statistical results derived from bootstrap samples, specifically within the "synergy_score_statistics" and "response_statistics" dataframes.
To illustrate my concern, for the "synergy_score_statistics" dataframe, the confidence interval boundaries given by [Bliss_synergy_ci_left, Bliss_synergy_ci_right] do not coincide with what I expected from calculating [Bliss_synergy_mean - 1.96Bliss_synergy_sem, Bliss_synergy_mean + 1.96Bliss_synergy_sem]. This same pattern of discrepancy is observed with other algorithms' statistics as well.
Additionally, in the "response_statistics" dataframe produced by the CalculateSensitivity function, I noticed that response_mean seems to be the average of the replicated dose-response matrix rather than the mean of bootstrap samples. Correspondingly, the response_sd and response_sem are also seemingly not derived from bootstrap samples, which differs from the approach taken in the "synergy_score_statistics" dataframe. I am curious as to why bootstrap means are not used for response_statistics as they are for synergy_score_statistics. Moreover, the confidence interval boundaries [response_ci_left, response_ci_right] are incongruous with the calculated boundaries [response_mean ± 1.96*response_sem].
I'm still honing my statistical understanding and would greatly appreciate clarification on these matters.
Thank you very much for your time, and I look forward to your insights. Here is the code follow the instruction:
`data("ONEIL_screening_data")
res <- ReshapeData(
data = ONEIL_screening_data,
data_type = "inhibition",
impute = TRUE,
impute_method = NULL,
noise = TRUE,
iteration = 10, # Number of iterations for bootstrapping
seed = 1
)
synergy <- CalculateSynergy(
data = res,
method = c("ZIP", "HSA", "Bliss", "Loewe"),
Emin = NA,
Emax = NA,
adjusted = TRUE,
correct_baseline = "non",
iteration = 10,
seed = 123
)
sensitivity <- CalculateSensitivity(
data = res,
correct_baseline = "non",
iteration = 10 # Number of iterations for bootstrapping
)`