Recompute regression line in log scale (#11)

- prettify scores (368527a4c9b600e8c64388f23ce01c9809d31d93)
- wip: explain scores (d231574b738e094431572bacda94a121568520cb)
- recompute regression line with log scale (bf971abc373ed1566f30112474ff0dc32235a377)

folding_studio_demo/app.py

@@ -300,14 +300,14 @@ def create_correlation_tab():
                     outputs=score_description,
                 )
         with gr.Column():
-            correlation_plot = gr.Plot(label="Correlation with binding affinity")
+            regression_plot = gr.Plot(label="Correlation with binding affinity")
 
     fake_predict_btn.click(
         fn=lambda x: fake_predict_and_correlate(
             spr_data_with_scores, SCORE_COLUMNS, ["Antibody Name", "KD (nM)"]
         ),
         inputs=[correlation_type],
-        outputs=[prediction_dataframe, correlation_ranking_plot, correlation_plot],
+        outputs=[prediction_dataframe, correlation_ranking_plot, regression_plot],
     )
 
     def update_regression_plot(score, use_log):
@@ -322,7 +322,7 @@ def create_correlation_tab():
     correlation_column.change(
         fn=update_regression_plot,
         inputs=[correlation_column, log_scale],
-        outputs=correlation_plot,
+        outputs=regression_plot,
     )
 
     correlation_type.change(
@@ -333,8 +333,8 @@ def create_correlation_tab():
 
     log_scale.change(
         fn=update_regression_plot,
-        inputs=[correlation_column, log_scale],
-        outputs=correlation_plot,
+        inputs=[correlation_column, log_scale], 
+        outputs=regression_plot,
     )
 
 

folding_studio_demo/correlate.py

@@ -157,9 +157,22 @@ def make_regression_plot(spr_data_with_scores: pd.DataFrame, score: str, use_log
         xaxis_type="log" if use_log else "linear"  # Set x-axis to logarithmic scale
     )
     # compute the regression line
-    corr_line = np.polyfit(spr_data_with_scores["KD (nM)"], spr_data_with_scores[score], 1)
-    corr_line_x = np.linspace(min(spr_data_with_scores["KD (nM)"]), max(spr_data_with_scores["KD (nM)"]), 100)
+    if use_log:
+        # Take log of KD values for fitting
+        x_vals = np.log10(spr_data_with_scores["KD (nM)"])
+    else:
+        x_vals = spr_data_with_scores["KD (nM)"]
+    
+    # Fit line to data
+    corr_line = np.polyfit(x_vals, spr_data_with_scores[score], 1)
+    
+    # Generate x points for line
+    corr_line_x = np.linspace(min(x_vals), max(x_vals), 100)
     corr_line_y = corr_line[0] * corr_line_x + corr_line[1]
+    
+    # Convert back from log space if needed
+    if use_log:
+        corr_line_x = 10**corr_line_x
     # add the regression line to the plot
     corr_plot.add_trace(go.Scatter(
         x=corr_line_x,