combine_data

Read in multiple data files and combine the effect and pvalue columns using specified functions (defaults are additive mean for effect and log mean for pvalue).

Parameters:

Name	Type	Description	Default
data_paths	List[str]	List of data file paths	required
identifier_col	str	Name of the feature identifier column	required
effect_col	str	Name of the effect column	required
pval_col	str	Name of the pvalue column	required
source	str	Source of the data	required
data_type	str	Type of data, either ‘binding’ or ‘expression’	required
combine_effect_func	Callable	Function to combine effect columns	mean
combine_pval_func	Callable	Function to combine pvalue columns	combine_pvals_detect_logged

Returns:

Type	Description
DataFrame	pd.DataFrame: Combined dataframe with averaged effect and pvalue

Source code in callingcardstools/Analysis/yeast/read_in_data.py

def combine_data(
    data_paths: List[str],
    identifier_col: str,
    effect_col: str,
    pval_col: str,
    source: str,
    data_type: Literal["binding", "expression"],
    combine_effect_func: Callable[[pd.Series], float] = np.mean,
    combine_pval_func: Callable[[pd.Series], float] = combine_pvals_detect_logged,
) -> pd.DataFrame:
    """
    Read in multiple data files and combine the effect and pvalue columns
    using specified functions (defaults are additive mean for effect and
    log mean for pvalue).

    Args:
        data_paths (List[str]): List of data file paths
        identifier_col (str): Name of the feature identifier column
        effect_col (str): Name of the effect column
        pval_col (str): Name of the pvalue column
        source (str): Source of the data
        data_type (str): Type of data, either 'binding' or 'expression'
        combine_effect_func (Callable): Function to combine effect columns
        combine_pval_func (Callable): Function to combine pvalue columns

    Returns:
        pd.DataFrame: Combined dataframe with averaged effect and pvalue
    """
    logger.info("combining data for data type {data_type} from {data_paths}")
    all_dfs = []

    for data_path in data_paths:
        df = read_in_data(
            data_path=data_path,
            identifier_col=identifier_col,
            effect_col=effect_col,
            pval_col=pval_col,
            source=source,
            data_type=data_type,
        )
        all_dfs.append(df)

    combined_df = (
        pd.concat(all_dfs)
        .groupby(["feature", f"{data_type}_source"])
        .agg(
            {
                f"{data_type}_effect": combine_effect_func,
                f"{data_type}_pvalue": combine_pval_func,
            }
        )
        .reset_index()
    )

    return combined_df