`epistasis.pyplot`¶

Optional matplotlib-backed plotting. Requires epistasis-v2[plot]. See Plotting epistatic coefficients for a guide-style walkthrough.

from epistasis.pyplot import plot_coefs, plot_correlation, DEFAULT_ORDER_COLORS

`plot_coefs`¶

def plot_coefs(
    model: EpistasisBaseModel | None = None,
    *,
    sites: Sequence[Site] | None = None,
    values: NDArray[np.float64] | None = None,
    stdeviations: NDArray[np.float64] | None = None,
    order_colors: Sequence[str] | None = None,
    sigmas: float = 0.0,
    significance: str | None = "bon",
    significance_cutoff: float = 0.05,
    star_cutoffs: tuple[float, ...] = (0.05, 0.01, 0.001),
    y_axis_name: str = "coefficient value",
    figsize: tuple[float, float] = (8.0, 5.0),
    height_ratio: float = 3.0,
    xgrid: bool = True,
    gridlines: float = 1.0,
    ax: list[Axes] | None = None,
) -> tuple[Figure, list[Axes]]

Plot epistatic coefficients as bars with a site-participation grid. Each bar is one coefficient, colored by interaction order; the grid underneath marks which sites participate in each term. The intercept term is dropped.

Coefficient source¶

Pass model (a fitted epistasis model) to read coefficients from its .epistasis map (sites, values, stdeviations).
Or pass sites and values directly. sites is a sequence of 1-indexed site tuples; the intercept is (0,).

These are mutually exclusive; supplying neither raises ValueError.

Key behaviors¶

Bars are colored by interaction order via order_colors (index 0 is the intercept / insignificant color, defaulting to DEFAULT_ORDER_COLORS). Vertical dotted lines separate orders.
The grid (xgrid=True) is one row per site, one column per coefficient; a cell is filled in that term's order color when the site participates.
When sigmas > 0 and standard errors are present, error bars are drawn, non-significant terms are greyed, and * stars are stacked per star_cutoffs threshold crossed. significance is "bon" (Bonferroni), "p" (raw), or None.

Returns¶

(fig, axes) where axes is [bar_axis, grid_axis] when xgrid is True, otherwise [bar_axis]. Pass ax to draw into existing axes instead of creating a new figure.

`plot_correlation`¶

def plot_correlation(
    model: EpistasisBaseModel | None = None,
    *,
    observed: NDArray[np.float64] | None = None,
    predicted: NDArray[np.float64] | None = None,
    color: str | None = None,
    point_size: float = 36.0,
    alpha: float = 0.85,
    annotate_r2: bool = True,
    figsize: tuple[float, float] = (5.5, 5.5),
    ax: Axes | None = None,
) -> tuple[Figure, Axes]

Scatter observed against predicted phenotype around the 1:1 line.

With model, observed values come from its attached GPM and predictions from model.predict().
With observed/predicted arrays, those are used directly.

The annotated R^2 is computed from the observed and predicted values.

Returns (fig, ax).

`DEFAULT_ORDER_COLORS`¶

A list of hex colors indexed by interaction order. Index 0 is reserved for the intercept and insignificant terms (grey); indices 1.. are the per-order colors. The defaults read on both light and dark backgrounds.

epistasis.pyplot¶

plot_coefs¶