cor_de: Conduct a differential expression analysis using correlation...
In mirvie/mirmodels: Models Built by the Informatics Team at Mirvie

Description Usage Arguments Details Value See Also Examples

This function is designed to ingest data in the form output by functions like get_bw_data().

cor_de(
  data,
  condition,
  genes = NULL,
  method = c("spearman", "kendall", "pearson"),
  padj_method = stats::p.adjust.methods[1]
)

`data`	A data frame where rows are samples and columns are genes and metadata.
`condition`	A string. The name of the column in `data` which contains the condition on which the differential expression is premised. This column should be logical, a factor with two levels or a numeric vector. The first level (or `FALSE`) is the baseline and the second level (or `TRUE`) is the disease/treatment. If the condition is numeric, then columns `log2fc`, `base_med`, `case_med`, `base_mean` and `case_mean` will be absent from the result.
`genes`	A character vector. The column names that contain the genes for the differential expression. If `NULL`, any genes in data found in `mirmisc::get_gene_names()` are used.
`method`	A string. The correlation test method. Must be `"spearman"`, `"kendall"` or `"pearson"`.
`padj_method`	A string. The method of adjusting the p-values for multiple hypothesis testing. Must be one of `stats::p.adjust.methods()`.

This function has no way of dealing with batch effects. To allow for these, you should correct your data (with e.g. linear_correct()) beforehand.

A tibble with 8 columns:

gene: The gene name.
log2fc: The log2 fold-change between case and control.
pvalue: The p-value for that gene being differentially expressed.
padj: Adjusted p-values.
base_med: The median raw value for the baseline samples.
case_med: The median raw value for the case samples.
base_mean: The mean raw value for the baseline samples.
case_mean: The mean raw value for the case samples. For method = "pearson" the log2fc is of the means of the two groups. Otherwise, it is of the medians.

Other differential expression methods: deseq(), edger()

if (requireNamespace("mirmisc", quietly = TRUE)) {
  bwms_data <- get_combined_cohort_data(
    c("bw", "ms"),
    log2 = TRUE, tot_counts = TRUE,
    gene_predicate = ~ stats::median(.) > 0,
  ) %>%
    dplyr::filter(!is.na(meta_pre_eclampsia)) %>%
    dplyr::mutate(tot_counts = log2(tot_counts + 1))
  genes <- dplyr::intersect(mirmisc::get_gene_names(), names(bwms_data))
  bwms_corr <- linear_correct(
    bwms_data,
    correct_cols = genes,
    correct_for_cols = c("log2_tot_counts", "cohort",
                         "meta_q_pcr_actb_ct", "meta_q_pcr_ercc_ct"),
    keep_effect_cols = "meta_pre_eclampsia"
  )[[1]]
  cor_de(bwms_corr, "meta_pre_eclampsia")
}