docs/expected_data.md

Expected data structure

This app has certain expectations towards the input data. As shown by the (constant) variable of module_input.R:

expected_data <- c(
    "dataMerged",
    "all_results",
    "configuration",
    "contrasteList",
    "rld",
    "txi.rsem"
)

the app expects precisely 6 variables (in a list, in a rds file), named this precise way, and with coherent sample/condition naming across all. They respectively contain: Important : the input data frame must not contain empty strings, only NAs! For instance, it might be the case that there is no gene name : in this case, the corresponding cell must contain NA. Below a simple code snippet to replace empty strings with NAs across a data frame : r dataframe_with_NAs <- dataframe_with_empty_str %>% mutate(across(everything(), na_if, ""))

dataMerged

A normalized matrix of gene counts. Posesses a column containing the genes' names, the column's name containing the word "symbol".

all_results

Named list (list) (with names of the type Condition_vs_Control) of the results dataframes, containing DESeq2 results (columns baseMean, padj, log2FoldChange, etc...), with a Row.names column containing the ENSEMBL genes IDs, and a (single) column of genes names which title contains the word "symbol". Ideally, also contains normalised expression values by sample. Expects in particular : - Row.names (Ensembl Gene IDS) - baseMean - log2FoldChange - padj - symbol (or a single column containing the word symbol, corresponding to gene names) - description

configuration

A dataframe with a column named File, another Name with the names of the samples, and a last "Condition" one with the coressponding conditions. The app does not support multiple conditions (what would be a multi-variable DESeq2 analysis, ~ Treatment + Cell type for instance). File is necessary as it is removed by named in some places.

contrasteList

Never actually used, as the app rather looks at the names of the all_results list to find the contrasts.

rld

DESeqTransform, result of one of the two (rlog or vst) variance stabilizing transformation.

txi.rsem

Result of the (tx)import of the results table, thus non-normalized.



GENOM-IC-Cochin/shiny-rnaseq-viz documentation built on Sept. 8, 2023, 4:23 p.m.