Q (haiku): Is KRAS G12C more aggressive than G12D?

▶ read_guide
{
 "uri": "cbioportal://statistical-tests-guide"
}

▶ read_guide
{
 "uri": "cbioportal://common-pitfalls"
}

◀ result
Statistical Test Selection Guide
================================

Purpose
-------
This guide ensures the correct statistical test is selected before performing any group comparison, matching cBioPortal's own Group Comparison defaults.

HARD RULES — NEVER FABRICATE A STATISTIC
----------------------------------------
ClickHouse cannot run statistical tests. The agent therefore must NEVER produce a derived statistic that is not a literal column value from a SQL result. Specifically:

1. **Never invent a p-value.** Not "p < 0.001", not "p ≈ 0.05", not any p-value. If the user asks "what is the p-value?", the answer is *"I can't compute that — here is the 2x2 contingency table (or group statistics). Run it in cBioPortal's Group Comparison tab, in R with `fisher.test(...)` / `wilcox.test(...)`, or in Python with `scipy.stats.fisher_exact(...)` / `mannwhitneyu(...)`."*
2. **Never claim mutual exclusivity (or co-occurrence) from a contingency table alone.** A 2x2 table is not a test. The shape "altered/not altered × group A/group B" needs Fisher's exact + a defined direction (odds ratio < 1 with significant p). Without that test, the agent presents the table and stops. Descriptive phrasing is also a claim: "largely/mostly mutually exclusive", "rarely co-occur", "tend to co-occur" are forbidden without the test — point to cBioPortal's Mutual Exclusivity tab instead.
3. **Never report a "median" that came from `AVG(...)` or any non-median aggregate.** "Median" and "mean" are different statistics; for skewed clinical distributions (especially survival) they differ substantially. Use ClickHouse's `quantile(0.5)(...)` for actual median, and label arithmetic averages as "mean", never "median".
4. **Never report a hazard ratio, odds ratio, risk ratio, or relative risk** that wasn't computed by an external tool. These require regression / model fitting that ClickHouse does not do.
5. **Never report median overall survival from `AVG(OS_MONTHS)` or even `quantile(0.5)(OS_MONTHS)`.** Median OS requires Kaplan-Meier estimation, which handles censoring (`OS_STATUS = 0:LIVING` means the event hasn't happened yet). Naive medians/means over `OS_MONTHS` ignore censoring and are systematically wrong. The correct handoff: return the raw `(OS_MONTHS, OS_STATUS)` pairs (or descriptive counts: N events, N censored, follow-up range) and tell the user to run KM in R (`survival::survfit`) or Python (`lifelines.KaplanMeierFitter`), or use cBioPortal's Survival comparison.

If asked to do any of (1)–(5), respond with the appropriate handoff template from the "Approved Response Templates" section below — do not produce the number.

Ambiguous-Term Clarification (before any analysis)
--------------------------------------------------
If the user's question uses an ambiguous term, ask before computing anything. Examples:

- **"aggressive"** → could mean shorter overall survival, higher metastasis rate, higher grade/stage, higher TMB, or specific molecular features. Ask which definit … (9984 more chars)

◀ result
# Common Query Pitfalls Guide

## Overview
This guide highlights frequent mistakes when analyzing cBioPortal data and provides solutions to avoid them.

## Critical Pitfalls

### 1. 🚨 CRITICAL MUTATION FREQUENCY ERRORS

#### ❌ WRONG: Using study-wide totals for gene frequencies
```sql
-- INCORRECT - This gives wrong frequencies!
SELECT
    hugo_gene_symbol,
    COUNT(DISTINCT sample_unique_id) as altered_samples,
    (SELECT COUNT(DISTINCT sample_unique_id)
     FROM genomic_event_derived
     WHERE cancer_study_identifier = 'your_study_id') as total_samples
FROM genomic_event_derived
WHERE variant_type = 'mutation' AND cancer_study_identifier = 'your_study_id'
GROUP BY hugo_gene_symbol;
```
**Problem**: Different genes have different profiling coverage - you can't use study-wide totals!

#### ❌ WRONG: Not using gene-specific profiling denominators
```sql
-- INCORRECT - Missing gene-specific denominators
SELECT
    hugo_gene_symbol,
    COUNT(DISTINCT sample_unique_id) as altered_samples
FROM genomic_event_derived
WHERE variant_type = 'mutation'
GROUP BY hugo_gene_symbol;
-- Missing: WHERE ARE THE DENOMINATORS FOR EACH GENE?
```

#### ❌ WRONG: Skipping individual gene profiling queries
**Problem**: Failing to run separate profiling queries for EACH gene in results.
**Each gene has different coverage**: TP53 might be profiled in 25,040 samples, MUC16 in 23,000, etc.

#### ✅ CORRECT: Complete gene-specific workflow
```sql
-- STEP 1: Get altered counts per gene
SELECT
    hugo_gene_symbol,
    entrez_gene_id,
    COUNT(DISTINCT CASE WHEN off_panel = 0 THEN sample_unique_id END) AS numberOfAlteredSamplesOnPanel,
    COUNT(*) AS totalMutationEvents
FROM genomic_event_derived
WHERE variant_type = 'mutation' AND mutation_status != 'UNCALLED'
GROUP BY entrez_gene_id, hugo_gene_symbol
ORDER BY numberOfAlteredSamplesOnPanel DESC;

-- STEP 2: FOR EACH GENE, run this profiling query:
SELECT COUNT(DISTINCT stgp.sample_unique_id) AS numberOfProfiledSamples
FROM sample_to_gene_panel_derived stgp
JOIN gene_panel gp ON stgp.gene_panel_id = gp.stable_id
JOIN gene_panel_list gpl ON gp.internal_id = gpl.internal_id
JOIN gene g ON gpl.gene_id = g.entrez_gene_id
WHERE stgp.alteration_type = 'MUTATION_EXTENDED'
  AND g.hugo_gene_symbol = 'TP53'  -- Replace with each gene from Step 1
  AND stgp.cancer_study_identifier = 'your_study_id';

-- STEP 3: Calculate frequency = numberOfAlteredSamplesOnPanel / numberOfProfiledSamples * 100
```

#### 🚨 WORKFLOW REQUIREMENTS VIOLATIONS:
- **Missing denominator columns**: Must show gene-specific profiled samples per row
- **Wrong table format**: Should be | Gene | # Mutations | # Samples | Profiled Samples | Sample % |
- **Not replacing gene symbols**: Must replace 'TP53' with actual gene for each query
- **Using study totals**: Never use genomic_event_derived for total sample counts

### 2. 🚨 OFF-PANEL MUTATION INCLUSION

#### ❌ Wrong: Including off-panel mutations
```sql
-- INCORRECT - Includes mutations outside gene panels
SELEC … (38012 more chars)

═ answer (success)
Thank you for this question. I need to **clarify what you mean by "aggressive"** before I can provide accurate data.

"Aggressive" could mean several things:
- **Overall survival (OS)** — which variant has shorter median survival?
- **Progression-free survival (PFS)** or **disease-free survival (DFS)**?
- **Grade or stage at diagnosis** — are G12C tumors higher stage?
- **Mutation frequency** — which is more common in advanced disease?
- **Clinical outcome** — response to therapy, time to recurrence?

Once you clarify, I'll pull the raw data from cBioPortal and point you to the right statistical test or visualization.

For **survival comparison** specifically, I would need to:
1. Filter samples by KRAS G12C vs G12D mutation status
2. Extract their OS_MONTHS and OS_STATUS values (to account for censoring)
3. Hand off to Kaplan-Meier estimation (cBioPortal's Group Comparison → Survival tab, or R/Python)

**I cannot produce a "median OS" or p-value myself** — ClickHouse cannot run statistical tests. But I can build the raw survival data and show you how to compute it.

**Which cancer type(s)** are you interested in? (e.g., lung adenocarcinoma, pancreatic cancer, colorectal cancer)