Regenerate all examples + notebooks after ref-allele off-by-one fix

The ref-allele correctness fix (commit 816a967) shifted predictions 1 bp,
so every committed example output was 1 bp off. This regenerates the full
cascade of walkthrough outputs and re-executes all three Python notebooks
from a clean kernel.

Walkthroughs regenerated (JSON + MD + TSV + HTML):
- variant_analysis: SORT1 AG, SORT1 ChromBPNet, SORT1 Enformer, BCL11A, FTO
- batch_scoring: SORT1 locus batch
- causal_prioritization: SORT1 locus
- sequence_engineering: region_swap, integration_simulation
- validation: SORT1 rs12740374 (AG solo, CEBP, multi-oracle), TERT
- discovery: SORT1 cell-type screen (472 cell types, top-3 per-CT reports)

Multi-oracle unified report regenerated with fresh per-oracle JSON + HTML
stubs (ChromBPNet, LegNet, AlphaGenome) and consolidated IGV.

Discovery top-3 cell types shifted after the 1-bp fix (stale HTML
deleted; new top-3 per-cell-type HTMLs written).

Notebooks re-executed inplace:
- single_oracle_quickstart: 7 plots, 0 errors
- advanced_multi_oracle_analysis: 57/57 cells, 16 plots, 0 errors
- comprehensive_oracle_showcase: 38/38 cells, 9 plots, 0 errors

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: lucapinello

examples/notebooks/advanced_multi_oracle_analysis.ipynb

@@ -41,7 +41,7 @@ <h3 style="margin:0 0 8px 0;font-size:1.05em;">Analysis Request</h3>
 <li><strong>Oracle:</strong> alphagenome</li>
 <li><strong>Normalizer:</strong> per-track background CDFs</li>
 <li><strong>Tracks requested:</strong> 6 HepG2 tracks</li>
-<li><strong>Generated:</strong> 2026-04-18 22:19 UTC</li>
+<li><strong>Generated:</strong> 2026-04-21 13:30 UTC</li>
 </ul>
 </section>
 <p><b>5 variants scored</b></p>
@@ -63,44 +63,44 @@ <h3 style="margin:0 0 8px 0;font-size:1.05em;">Analysis Request</h3>
 <th>Ref</th><th>Alt</th><th>Effect <span class="formula-chip">log2FC</span></th><th>Effect %ile</th>
 </tr></thead><tbody>
 <tr><td>chr1:109274968 G>T</td><td>rs12740374</td>
-<td>512</td><td>699</td><td class='gain'>+0.448</td><td>≥99th</td>
-<td>2.1e+03</td><td>2.72e+03</td><td class='gain'>+0.371</td><td>≥99th</td>
-<td>1.22e+03</td><td>1.46e+03</td><td class='gain'>+0.268</td><td>≥99th</td>
-<td>1.37e+04</td><td>1.55e+04</td><td class='gain'>+0.180</td><td>≥99th</td>
-<td>39</td><td>38.8</td><td class='neutral'>-0.004</td><td>≥99th</td>
-<td>3.85e+03</td><td>3.84e+03</td><td class='neutral'>-0.004</td><td>≥99th</td>
+<td>595</td><td>1.57e+03</td><td class='gain'>+1.400</td><td>≥99th</td>
+<td>2.54e+03</td><td>1.72e+04</td><td class='gain'>+2.760</td><td>≥99th</td>
+<td>1.4e+03</td><td>1.13e+04</td><td class='gain'>+3.020</td><td>≥99th</td>
+<td>1.43e+04</td><td>3.6e+04</td><td class='gain'>+1.336</td><td>≥99th</td>
+<td>39</td><td>38.8</td><td class='neutral'>-0.008</td><td>≥99th</td>
+<td>3.86e+03</td><td>3.85e+03</td><td class='neutral'>-0.004</td><td>≥99th</td>
 </tr>
-<tr><td>chr1:109279175 G>A</td><td>rs4970836</td>
-<td>8.22</td><td>7.96</td><td class='neutral'>-0.041</td><td>≥99th</td>
-<td>201</td><td>197</td><td class='neutral'>-0.031</td><td>≥99th</td>
-<td>202</td><td>198</td><td class='neutral'>-0.027</td><td>≥99th</td>
-<td>3.32e+03</td><td>3.28e+03</td><td class='neutral'>-0.015</td><td>≥99th</td>
-<td>43.2</td><td>43.3</td><td class='neutral'>+0.004</td><td>≥99th</td>
-<td>4.19e+03</td><td>4.18e+03</td><td class='neutral'>-0.001</td><td>≥99th</td>
+<tr><td>chr1:109274570 A>G</td><td>rs7528419</td>
+<td>119</td><td>113</td><td class='neutral'>-0.067</td><td>≥99th</td>
+<td>967</td><td>968</td><td class='neutral'>+0.001</td><td>near-zero</td>
+<td>771</td><td>766</td><td class='neutral'>-0.010</td><td>≥99th</td>
+<td>1.53e+04</td><td>1.5e+04</td><td class='neutral'>-0.025</td><td>≥99th</td>
+<td>42</td><td>42</td><td class='neutral'>-0.002</td><td>≥99th</td>
+<td>4.38e+03</td><td>4.37e+03</td><td class='neutral'>-0.003</td><td>≥99th</td>
 </tr>
-<tr><td>chr1:109275216 T>C</td><td>rs660240</td>
-<td>398</td><td>418</td><td class='neutral'>+0.071</td><td>≥99th</td>
-<td>1.28e+03</td><td>1.3e+03</td><td class='neutral'>+0.021</td><td>≥99th</td>
-<td>770</td><td>779</td><td class='neutral'>+0.017</td><td>≥99th</td>
-<td>1.63e+04</td><td>1.67e+04</td><td class='neutral'>+0.032</td><td>≥99th</td>
-<td>45</td><td>45.1</td><td class='neutral'>+0.003</td><td>≥99th</td>
-<td>4.31e+03</td><td>4.31e+03</td><td class='neutral'>-0.000</td><td>near-zero</td>
+<tr><td>chr1:109279175 G>A</td><td>rs4970836</td>
+<td>7.99</td><td>7.84</td><td class='neutral'>-0.025</td><td>≥99th</td>
+<td>197</td><td>196</td><td class='neutral'>-0.009</td><td>≥99th</td>
+<td>197</td><td>196</td><td class='neutral'>-0.007</td><td>≥99th</td>
+<td>3.29e+03</td><td>3.31e+03</td><td class='neutral'>+0.007</td><td>≥99th</td>
+<td>35.4</td><td>35.4</td><td class='neutral'>+0.003</td><td>≥99th</td>
+<td>3.94e+03</td><td>3.93e+03</td><td class='neutral'>-0.001</td><td>near-zero</td>
 </tr>
 <tr><td>chr1:109275684 G>T</td><td>rs1626484</td>
-<td>69.4</td><td>69.6</td><td class='neutral'>+0.003</td><td>0.67</td>
-<td>539</td><td>541</td><td class='neutral'>+0.005</td><td>≥99th</td>
-<td>512</td><td>512</td><td class='neutral'>+0.001</td><td>0.21</td>
-<td>1.34e+04</td><td>1.34e+04</td><td class='neutral'>-0.000</td><td>near-zero</td>
-<td>37.3</td><td>37.3</td><td class='neutral'>+0.000</td><td>near-zero</td>
-<td>3.94e+03</td><td>3.95e+03</td><td class='neutral'>+0.001</td><td>≥99th</td>
+<td>63.4</td><td>60.9</td><td class='neutral'>-0.057</td><td>≥99th</td>
+<td>527</td><td>522</td><td class='neutral'>-0.015</td><td>≥99th</td>
+<td>512</td><td>510</td><td class='neutral'>-0.004</td><td>0.78</td>
+<td>1.2e+04</td><td>1.19e+04</td><td class='neutral'>-0.016</td><td>≥99th</td>
+<td>41.8</td><td>42</td><td class='neutral'>+0.006</td><td>≥99th</td>
+<td>4.14e+03</td><td>4.14e+03</td><td class='neutral'>-0.001</td><td>near-zero</td>
 </tr>
-<tr><td>chr1:109274570 A>G</td><td>rs7528419</td>
-<td>118</td><td>119</td><td class='neutral'>+0.008</td><td>≥99th</td>
-<td>958</td><td>960</td><td class='neutral'>+0.002</td><td>0.62</td>
-<td>776</td><td>778</td><td class='neutral'>+0.002</td><td>0.31</td>
-<td>1.46e+04</td><td>1.48e+04</td><td class='neutral'>+0.018</td><td>≥99th</td>
-<td>40.7</td><td>40.7</td><td class='neutral'>-0.003</td><td>≥99th</td>
-<td>4.3e+03</td><td>4.29e+03</td><td class='neutral'>-0.002</td><td>≥99th</td>
+<tr><td>chr1:109275216 T>C</td><td>rs660240</td>
+<td>368</td><td>371</td><td class='neutral'>+0.014</td><td>≥99th</td>
+<td>1.27e+03</td><td>1.28e+03</td><td class='neutral'>+0.009</td><td>≥99th</td>
+<td>768</td><td>773</td><td class='neutral'>+0.009</td><td>≥99th</td>
+<td>1.56e+04</td><td>1.56e+04</td><td class='neutral'>-0.002</td><td>0.69</td>
+<td>42.1</td><td>41.7</td><td class='neutral'>-0.014</td><td>≥99th</td>
+<td>4.1e+03</td><td>4.09e+03</td><td class='neutral'>-0.006</td><td>≥99th</td>
 </tr>
 </tbody></table>
 <p style='font-size:0.85em;color:#6b7280;margin-top:8px;'>Columns per track: <b>Ref</b> (reference allele prediction), <b>Alt</b> (alternate allele prediction), <b>Effect</b> (signed effect — formula shown in the chip next to each track's header; see the glossary at the top of the page), <b>Effect %ile</b> (ranked against ~10K random SNPs). Green = gain, red = loss.</p>