Q (haiku): Which genomic alterations are associated with immunotherapy response in melanoma patients from the MSK-CHORD Study?

▶ read_guide
{
 "uri": "cbioportal://gene-expression-guide"
}

▶ list_studies
{
 "search": "MSK CHORD melanoma"
}

◀ result
# Gene Expression Analysis Guide

This guide covers continuous-value genomic data: gene **expression**, **copy number** values, **methylation**, and related profile types. Mutation/CNA/SV *frequency* analysis lives in `cbioportal://mutation-frequency-guide`.

## Where this data lives

Continuous per-sample-per-gene values are stored in `genetic_alteration_derived`:

| Column | Description |
|---|---|
| `sample_unique_id` | `<study>_<sample_stable_id>` |
| `cancer_study_identifier` | study scope |
| `hugo_gene_symbol` | gene |
| `profile_type` | which assay/normalization (see below) |
| `alteration_value` | the actual value — stored as Nullable(String); cast with `toFloat64OrNull` |

`alteration_value` is a string because the same column hosts many different value scales. The `''` and `'NA'` sentinels mean "missing"; always filter them out and use `toFloat64OrNull(alteration_value) IS NOT NULL` for downstream math.

## Discovering profile types for a study

Different studies expose different profile types depending on what assays were run and how the data was normalized. Always check what a specific study supports before picking one:

```sql
SELECT DISTINCT profile_type
FROM genetic_alteration_derived
WHERE cancer_study_identifier = 'brca_metabric'
ORDER BY profile_type;
```

Common values across the public portal:

| Family | Profile types |
|---|---|
| mRNA expression | `rna_seq_v2_mrna`, `rna_seq_v2_mrna_median_Zscores`, `rna_seq_v2_mrna_median_all_sample_Zscores` (TCGA PanCancer Atlas), `mrna`, `mrna_median_Zscores`, `mrna_seq_v2_rsem`, `mrna_seq_v2_rsem_Zscores`, `mrna_seq_cpm`, `mrna_seq_fpkm`, `mrna_U133`, `mrna_outliers` |
| Copy number (continuous) | `cna`, `linear_CNA`, `log2CNA`, `cna_consensus`, `cna_rae`, `gistic` |
| Methylation | `methylation_hm27`, `methylation_hm450`, `methylation_epic`, `methylation_promoters_rrbs` |
| miRNA | `mirna`, `mirna_median_Zscores` |
| Protein | `protein_quantification`, `protein_level`, `RPPA` |

**Z-score vs raw choice.** When the user asks "is X correlated with Y", either works for Spearman (rank-based) — Pearson would care. Default to the non-Z-score variant if both exist, and call out which one in the response.

## Canonical recipe — Spearman correlation between two genes

```sql
SELECT *
FROM gene_pair_coexpression(
    study        = 'brca_metabric',
    gene_a       = 'TP53',
    gene_b       = 'MYC',
    profile_type = 'mrna'
);
```

Returns one row: `(gene_a, gene_b, profile_type, spearman_correlation, num_samples)`.

- `spearman_correlation` in [−1, 1]; `NULL` when fewer than 3 valid paired samples.
- Mirrors cbioportal-backend's `ClickhouseCoExpressionMapper.getCoExpressions`, simplified to a pair lookup (the backend computes one ref gene vs ALL other genes for the coexpression page; here the agent asks about a specific pair).

### Verified examples

| Study | gene_a | gene_b | profile_type | spearman | n |
|---|---|---|---|---|---|
| `brca_metabric` | TP53 | MYC | `mrna` | 0.118 | 1980 |
|  … (4931 more chars)

◀ result
{"result":[]}

▶ list_studies
{
 "search": "CHORD"
}

◀ result
{"result":[{"cancer_study_identifier":"msk_chord_2024","name":"MSK-CHORD (MSK, Nature 2024)","type_of_cancer_id":"mixed","sample_count":25040,"has_guide":true,"url":"https://www.cbioportal.org/study/summary?id=msk_chord_2024"}]}

▶ get_study_guide
{
 "study_id": "msk_chord_2024"
}

▶ read_guide
{
 "uri": "cbioportal://treatment-guide"
}

◀ result
# MSK-CHORD (MSK, Nature 2024)

**Study ID:** `msk_chord_2024`

## Overview
Targeted sequencing via MSK-IMPACT panels. Clinical annotations include some derived from natural language processing (denoted NLP).

**Exactly five cancer types** (`CANCER_TYPE`, patients): Non-Small Cell Lung Cancer 7,809, Colorectal Cancer 5,543, Breast Cancer 5,368, Prostate Cancer 3,211, Pancreatic Cancer 3,109. There is **no melanoma** or any other cancer type; say so up front if asked, instead of substituting another type.

**No therapy-response variable.** There is no RECIST, objective response, or best-response attribute or event. For treatment-outcome questions (e.g. immunotherapy response), say this first; the only proxies are `OS_MONTHS`/`OS_STATUS`, or NLP radiology progression events (`Diagnosis` events with `SUBTYPE = 'Progression'`, key `PROGRESSION` = Y/N/Indeterminate), in patients with `Treatment` events of the relevant `SUBTYPE` (e.g. `Immuno`: 3,341 patients). Hand off the comparison to cBioPortal group comparison / survival.

**Nearly one sample per patient: 24,950 patients / 25,040 samples.** Only 90 patients have more than one sample, and all 90 have samples from two different cancer types (second primaries); only 26 have both a `Primary` and a `Metastasis` sample. There is no meaningful same-patient (paired) primary-vs-metastasis cohort. For "same patient" / paired questions, say this up front, then offer the **unpaired** comparison of all `Primary` vs `Metastasis` samples (`SAMPLE_TYPE`), labelled as unpaired.
```sql
SELECT countIf(n > 1) AS multi_sample_patients,         -- 90
       countIf(has_p AND has_m) AS primary_and_met      -- 26
FROM (SELECT patient_unique_id, count() AS n,
             has(groupArray(attribute_value), 'Primary') AS has_p,
             has(groupArray(attribute_value), 'Metastasis') AS has_m
      FROM clinical_data_derived
      WHERE cancer_study_identifier = 'msk_chord_2024' AND attribute_name = 'SAMPLE_TYPE'
      GROUP BY patient_unique_id);
```

## Gene Panels
This study uses multiple MSK-IMPACT panel versions:
- **IMPACT341**: Earlier version, 341 genes
- **IMPACT410**: 410 genes
- **IMPACT468**: 468 genes  
- **IMPACT505**: Latest version, 505 genes

**Important:** Different samples may have different gene coverage. Always use gene-specific denominators when calculating mutation frequencies.

## Clinical Attributes - Semantic Guide

### Cancer Classification
| Attribute | Description | Values |
|-----------|-------------|--------|
| `CANCER_TYPE` | Broad cancer category | e.g., "Non-Small Cell Lung Cancer", "Breast Cancer" |
| `CANCER_TYPE_DETAILED` | Specific subtype | e.g., "Lung Adenocarcinoma", "Invasive Ductal Carcinoma" |
| `ONCOTREE_CODE` | OncoTree classification code | Standardized cancer type codes |

### Sample Information
| Attribute | Description | Values |
|-----------|-------------|--------|
| `SAMPLE_TYPE` | Sample origin | Primary, Metastasis, Local Recurrence, Unknown |
| `SAMPLE_CLASS` | Sample … (5439 more chars)

▶ read_guide
{
 "uri": "cbioportal://external-resources-guide"
}

◀ result
# Treatment Data Query Guide

## Overview

Treatment data in cBioPortal is stored in **clinical event tables**, separate from clinical attributes. This allows for timeline-based treatment records with start/stop dates.

## Start Here: Treatment Views

For "most common treatments / agents / regimens in study X", use the parameterized views (documented in `cbioportal://clinical-data-guide`, Study-View Chart Counts):

```sql
-- Patients per agent (the portal's Treatment chart), with type/subtype arrays
SELECT * FROM treatment_counts_in_study(study='msk_chord_2024') ORDER BY patients DESC LIMIT 20;

-- Same-day agent combinations (investigational, prior-medication and radiation events excluded)
SELECT * FROM treatment_regimens_in_study(study='msk_chord_2024') ORDER BY patients DESC LIMIT 20;
```

Write raw event queries (below) only for subgroups, timelines or keys the views do not expose.

## Key Tables

| Table | Description |
|-------|-------------|
| `clinical_event` | Event records with patient_id, event_type, start_date, stop_date |
| `clinical_event_data` | Key-value pairs linked to each clinical_event_id |

## Schema

```
clinical_event
├── clinical_event_id (PK)
├── patient_id (FK → patient.internal_id)
├── event_type (Treatment, TREATMENT, Diagnosis, SURGERY, etc.)
├── start_date (days from diagnosis)
└── stop_date (days from diagnosis)

clinical_event_data
├── clinical_event_id (FK)
├── key (AGENT, SUBTYPE, etc.)
└── value
```

## Event Types

Not all studies have all event types. Common ones include:

| Event Type | Description |
|------------|-------------|
| `Treatment` / `TREATMENT` | Drug/therapy administration |
| `SURGERY` | Surgical procedures |
| `Diagnosis` | Diagnosis events |
| `LAB_TEST` | Laboratory results |
| `Sequencing` | Sequencing events |
| `Sample acquisition` | Sample collection |
| `PATHOLOGY` / `Pathology` | Pathology reports |

## Basic Treatment Queries

### List Available Event Types in a Study

```sql
SELECT DISTINCT ce.event_type, COUNT(*) as event_count
FROM clinical_event ce
JOIN patient p ON ce.patient_id = p.internal_id
JOIN cancer_study cs ON p.cancer_study_id = cs.cancer_study_id
WHERE cs.cancer_study_identifier = 'your_study_id'
GROUP BY ce.event_type
ORDER BY event_count DESC;
```

### List Treatment Data Keys in a Study

```sql
SELECT DISTINCT ced.key, COUNT(*) as cnt
FROM clinical_event ce
JOIN clinical_event_data ced ON ce.clinical_event_id = ced.clinical_event_id
JOIN patient p ON ce.patient_id = p.internal_id
JOIN cancer_study cs ON p.cancer_study_id = cs.cancer_study_id
WHERE cs.cancer_study_identifier = 'your_study_id'
  AND ce.event_type IN ('Treatment', 'TREATMENT')
GROUP BY ced.key
ORDER BY cnt DESC;
```

### Get Most Common Treatment Agents

```sql
SELECT 
    ced.value as agent,
    COUNT(DISTINCT ce.patient_id) as patient_count
FROM clinical_event ce
JOIN clinical_event_data ced ON ce.clinical_event_id = ced.clinical_event_id
JOIN patient p ON ce.patient_id = p.internal_id
JOIN cancer_stud … (6220 more chars)

◀ result
# External Resources Guide

Use this guide when the user asks about data that may be linked from cBioPortal rather than stored directly in molecular or clinical tables.

## Routing Triggers

Read this guide before answering or refusing questions that mention:

- imaging, radiology, CT, MRI, pathology slides, histology, Minerva, viewer
- external portal, external resource, image data, spatial data
- HTAN studies or study-specific linked viewers

## Core Rule

Do not say cBioPortal has no imaging or external-resource data until you have checked:

- `resource_definition`
- `resource_study`
- `resource_sample`
- `resource_patient`

cBioPortal may store links to external viewers or portals even when it does not store raw images.

## Which Studies Have Imaging Data

For "which studies have imaging / pathology slides / CT" or "how many samples have images", read `cancer_study.resource_sample_counts` — per-study sample counts keyed by resource display name, the same numbers as the portal's "Data type" filter. One query, no joins:

```sql
SELECT
    cancer_study_identifier,
    name,
    resource_sample_counts['Slide Microscopy'] AS slide_microscopy_samples
FROM cancer_study
WHERE resource_sample_counts['Slide Microscopy'] > 0
ORDER BY slide_microscopy_samples DESC;
```

List the resource names that exist with `SELECT DISTINCT arrayJoin(mapKeys(resource_sample_counts)) FROM cancer_study` (e.g. `'Slide Microscopy'`, `'Computed Tomography'`, `'Magnetic Resonance'`, `'H&E Slide'`, `'MxIF Image'`).

The map counts sample- and patient-level resources only. For study-level links (`resource_study`) and for the URLs themselves, use the queries below.

## Discovery Query

Start with table and column validation, then use this pattern:

```sql
SELECT
    rd.resource_id,
    rd.display_name,
    rd.description,
    rd.resource_type,
    rs.cancer_study_identifier,
    rs.url
FROM resource_study rs
JOIN resource_definition rd
    ON rs.resource_id = rd.resource_id
WHERE lower(rd.display_name) LIKE '%minerva%'
   OR lower(rd.description) LIKE '%minerva%'
   OR lower(rd.display_name) LIKE '%image%'
   OR lower(rd.description) LIKE '%image%'
   OR lower(rd.display_name) LIKE '%pathology%'
   OR lower(rd.description) LIKE '%pathology%'
   OR lower(rd.display_name) LIKE '%histology%'
   OR lower(rd.description) LIKE '%histology%'
ORDER BY rs.cancer_study_identifier, rd.display_name
LIMIT 100;
```

If no study-level rows appear, check sample- and patient-level resource links:

```sql
SELECT
    rd.resource_id,
    rd.display_name,
    rd.description,
    rs.cancer_study_identifier,
    rs.sample_unique_id,
    rs.url
FROM resource_sample rs
JOIN resource_definition rd
    ON rs.resource_id = rd.resource_id
WHERE lower(rd.display_name) LIKE '%minerva%'
   OR lower(rd.description) LIKE '%minerva%'
   OR lower(rd.display_name) LIKE '%image%'
   OR lower(rd.description) LIKE '%image%'
   OR lower(rd.display_name) LIKE '%pathology%'
   OR lower(rd.description) LIKE '%pathology%' … (981 more chars)

▶ clickhouse_run_select_query
{
 "query": "-- First, confirm melanoma is not in MSK-CHORD
SELECT DISTINCT attribute_value AS cancer_type, COUNT(DISTINCT sample_unique_id) as samples
FROM clinical_data_derived
WHERE cancer_study_identifier = 'msk_chord_2024' 
  AND attribute_name = 'CANCER_TYPE'
GROUP BY attribute_value
ORDER BY samples DESC;"
}

◀ result
{"rows":[{"cancer_type":"Non-Small Cell Lung Cancer","samples":7809},{"cancer_type":"Colorectal Cancer","samples":5543},{"cancer_type":"Breast Cancer","samples":5368},{"cancer_type":"Prostate Cancer","samples":3211},{"cancer_type":"Pancreatic Cancer","samples":3109}]}

▶ clickhouse_run_select_query
{
 "query": "-- Check for immunotherapy treatments in MSK-CHORD
SELECT 
    subtype.value as treatment_subtype,
    agent.value as agent_name,
    COUNT(DISTINCT ce.patient_id) as patients
FROM clinical_event ce
JOIN clinical_event_data agent ON ce.clinical_event_id = agent.clinical_event_id AND agent.key = 'AGENT'
JOIN clinical_event_data subtype ON ce.clinical_event_id = subtype.clinical_event_id AND subtype.key = 'SUBTYPE'
JOIN patient p ON ce.patient_id = p.internal_id
JOIN cancer_study cs ON p.cancer_study_id = cs.cancer_study_id
WHERE cs.cancer_study_identifier = 'msk_chord_2024'
  AND ce.event_type IN ('Treatment', 'TREATMENT')
  AND subtype.value = 'Immuno'
GROUP BY subtype.value, agent.value
ORDER BY patients DESC
LIMIT 20;"
}

◀ result
{"rows":[{"treatment_subtype":"Immuno","agent_name":"PEMBROLIZUMAB","patients":2061},{"treatment_subtype":"Immuno","agent_name":"NIVOLUMAB","patients":809},{"treatment_subtype":"Immuno","agent_name":"ATEZOLIZUMAB","patients":370},{"treatment_subtype":"Immuno","agent_name":"DURVALUMAB","patients":246},{"treatment_subtype":"Immuno","agent_name":"IPILIMUMAB","patients":218},{"treatment_subtype":"Immuno","agent_name":"AVELUMAB","patients":3},{"treatment_subtype":"Immuno","agent_name":"CEMIPLIMAB","patients":3},{"treatment_subtype":"Immuno","agent_name":"TREMELIMUMAB","patients":1}]}

═ answer (success)
Based on my investigation, I need to provide you with an important clarification:

## Key Finding: Melanoma is Not in MSK-CHORD

**The MSK-CHORD study (MSK, Nature 2024) does not contain melanoma patients.** The cohort consists of exactly five cancer types:
- Non-Small Cell Lung Cancer (7,809 samples)
- Colorectal Cancer (5,543 samples)
- Breast Cancer (5,368 samples)
- Prostate Cancer (3,211 samples)
- Pancreatic Cancer (3,109 samples)

---

## Important Data Limitation: No Therapy Response Phenotype

**MSK-CHORD does NOT include immunotherapy response data** (e.g., RECIST response, objective response rate, or best response). According to the study guide, there is no therapy-response variable in the clinical attributes or events.

---

## What IS Available for Immunotherapy in MSK-CHORD

While immunotherapy response is not measured, the study does have:

1. **Immunotherapy Treatment Records**: 3,341 patients across all cancer types received immunotherapy, with top agents:
   - **Pembrolizumab**: 2,061 patients
   - **Nivolumab**: 809 patients
   - **Atezolizumab**: 370 patients
   - **Durvalumab**: 246 patients
   - **Ipilimumab**: 218 patients

2. **Genomic Alterations**: Sequenced via MSK-IMPACT panels (341–505 genes depending on panel version)

3. **Potential Outcome Proxies**: 
   - Overall survival (`OS_MONTHS`, `OS_STATUS`)
   - NLP-derived progression events (`PROGRESSION = Y/N`)

---

## What You Can Do Instead

If you're interested in **genomic alterations and immunotherapy in lung or colorectal cancer** (which both have substantial immunotherapy use in MSK-CHORD), I can:

1. **Identify frequently mutated genes** in immunotherapy-treated patients by cancer type
2. **Link genomic alterations to survival outcomes** (using OS/progression as a proxy)
3. **Compare genomic profiles** between immunotherapy-treated and untreated patients

**For melanoma immunotherapy genomics**, you would need a different study—[cBioPortal hosts several skin cancer cohorts](https://www.cbioportal.org). Would you like me to search for melanoma studies with immunotherapy data instead?