feat: furtehr analysis added

notebooks/initial_eda_lb_20250919.py

@@ -840,6 +840,141 @@ def _(sc, variants_high_cov_ids, variants_lost_cov_ids, variants_low_cov_ids):
     return
 
 
+@app.cell
+def _(mo):
+    mo.md(r"""# (3) Stats and Metadata""")
+    return
+
+
+@app.cell
+def _(pd):
+    meta = pd.read_excel(
+        "/home/darren/Documents/4_data/3_internal/2_lb/Metadata_LB_20250924.xlsx",
+        index_col="fastq_id",
+    )
+    meta
+    return (meta,)
+
+
+@app.cell
+def _(clinicals, meta, pd, stats):
+    # merge stats and metadata and remove seracares
+    stats_meta = pd.concat([stats, meta], axis=1)
+    stats_meta = stats_meta.loc[clinicals, :]
+    stats_meta = stats_meta.reset_index()
+    stats_meta.head()
+    return (stats_meta,)
+
+
+@app.cell
+def _(mo, stats_meta):
+    var1 = mo.ui.dropdown(options=stats_meta.columns, label="Variable 1")
+    var2 = mo.ui.dropdown(options=stats_meta.columns, label="Variable 2")
+
+    mo.vstack([var1, var2])
+    return var1, var2
+
+
+@app.cell
+def _(px, stats_meta, var1, var2):
+    # note that all SE1 clinical samples are older QG extracted cfDNAs
+    fig5 = px.scatter(
+        data_frame=stats_meta,
+        x=var1.value,
+        y=var2.value,
+        color="Flow cell #",
+        width=700,
+        height=500,
+        template="simple_white",
+        hover_data=["index"],
+        color_discrete_sequence=px.colors.qualitative.Dark2,
+        category_orders={"Flow cell #": ["SE1", "SE2"]},
+        trendline="ols",
+    )
+    fig5.update_traces(
+        marker=dict(size=9, line=dict(width=1, color="DarkSlateGrey")),
+        selector=dict(mode="markers"),
+    )
+    fig5.add_vline(
+        x=30,
+        line_width=3,
+        line_dash="dot",
+        annotation_text="30ng",
+        annotation_position="top left",
+    )
+    fig5.show()
+    return
+
+
+@app.cell
+def _(stats_meta):
+    stats_meta["ng_group"] = [
+        "< 30ng" if i < 30 else "> 30ng" for i in stats_meta["Input DNA used (ng)"]
+    ]
+    return
+
+
+@app.cell
+def _(pd, stats_meta):
+    stats_meta_melt = pd.melt(
+        stats_meta, id_vars=["index", "sample_type", "ng_group", "Flow cell #"]
+    )
+    stats_meta_melt.head()
+    return (stats_meta_melt,)
+
+
+@app.cell
+def _(mo, stats_meta_melt):
+    box_vars = mo.ui.dropdown(
+        options=stats_meta_melt["variable"].unique(), label="Variables to plot"
+    )
+    mo.hstack([box_vars])
+    return (box_vars,)
+
+
+@app.cell
+def _(box_vars, px, stats_meta_melt):
+    fig6 = px.box(
+        data_frame=stats_meta_melt.loc[
+            (stats_meta_melt["variable"] == box_vars.value)
+            & (stats_meta_melt["Flow cell #"] == "SE2")
+        ],
+        color="ng_group",
+        y="value",
+        x="variable",
+        template="simple_white",
+        points="all",
+        labels={
+            "ng_group": "cfDNA Input for Library",
+            "value": box_vars.value,
+            "variable": "",
+        },
+        color_discrete_sequence=px.colors.qualitative.Dark2,
+        category_orders={"ng_group": ["< 30ng", "> 30ng"]},
+        hover_data=["index"],
+        width=800,
+    )
+    fig6.update_traces(
+        marker=dict(size=10, line=dict(width=1, color="DarkSlateGrey")),
+        selector=dict(type="points"),
+    )
+    fig6.update_xaxes(ticks="", showticklabels=False)
+    fig6.update_yaxes(showgrid=True, tickfont_size=14)
+    fig6.update_legends(font_size=14)
+
+    for trace in fig6.select_traces():
+        trace.marker.update(size=10, line=dict(width=1, color="DarkSlateGrey"))
+
+    fig6.show()
+    return
+
+
+@app.cell
+def _(stats_meta):
+    stats_meta["ng_group"].value_counts()
+    return
+
+
 @app.cell
 def _():
     return