# ROS1 dependency and drugs in one BioQ query

> ROS1 dependency and drugs from one query: 0 of 1178 DepMap lines dependent, 5 drugs in ChEMBL and Open Targets, 711 disease links, as of 2026-11-06.

Data as of 2026-11-06 · Published 2026-11-10 · Updated 2026-11-06 · https://bioq.bio/blog/en/gene-ros1-profile

![BioQ result page for the query ROS1, showing the CRISPR dependency summary, the drug list and the disease association table in one view](https://bioq.bio/blog/img/gene-ros1-profile/en/hero.webp)

For ROS1 dependency and drugs, one BioQ query shows that none of 1178 DepMap cell lines count as dependent on ROS1 (mean Chronos gene effect 0.094). Meanwhile 5 drugs are linked to the gene, and Open Targets lists 711 associated diseases, led by non-small cell lung carcinoma at a score of 0.713. The data are as of 2026-11-06. The layers disagree on the surface: no pan-cancer dependency, yet strong drug and clinical evidence in lung cancer.

## What we asked

The query was the bare gene symbol, ROS1. We wanted one view of a kinase that is best known as a fusion-driven oncogene, across five questions:

- Does CRISPR knockout reduce viability across cell lines?
- Which drugs act on it, and how potent are they?
- Which diseases carry the most evidence?
- Is the gene expressed differently in tumour and normal tissue?
- What do variant-level clinical records and interaction partners add?

BioQ ran 7 engine queries across 6 distinct resources in 0.59 engine seconds. Done by hand, the same questions mean visiting each portal, matching gene identifiers and reading each release note. The time saving is mostly in that assembly step. You still have to read the results critically.

## What BioQ returned for ROS1 dependency and drugs

**Dependency.** In DepMap (release depmap_24q4, Chronos), 0 of 1178 cell lines pass the dependency threshold. The mean gene effect is 0.094, and ROS1 is not flagged as common essential. A more negative Chronos value means a stronger viability loss, so a small positive mean indicates no knockout cost on average. The lineage rows below are the eight shown.

| Lineage | Cell lines | Dependent | % dependent | Mean gene effect |
|---|---|---|---|---|
| Cervix | 18 | 0 | 0 | 0.072 |
| Testis | 5 | 0 | 0 | 0.073 |
| Lung | 126 | 0 | 0 | 0.076 |
| Prostate | 10 | 0 | 0 | 0.08 |
| Liver | 24 | 0 | 0 | 0.086 |
| Uterus | 34 | 0 | 0 | 0.086 |
| Pleura | 21 | 0 | 0 | 0.087 |
| Lymphoid | 93 | 0 | 0 | 0.087 |

![BioQ DepMap panel for ROS1 with Chronos gene effect by lineage, showing no dependent cell lines](https://bioq.bio/blog/img/gene-ros1-profile/en/dep.webp)
*DepMap Chronos view for ROS1 in BioQ, split by lineage.*

**Drugs.** Five drugs are listed. Four have a maximum clinical stage of APPROVAL, and taletrectinib is at PHASE_3.

| Drug | Max clinical stage | Mechanism | Best exact pChEMBL | Type | Trials by name |
|---|---|---|---|---|---|
| Repotrectinib | APPROVAL | Proto-oncogene tyrosine-protein kinase ROS inhibitor | 10.15 | IC50 | 12 |
| Lorlatinib | APPROVAL | — | 10 | KI | 51 |
| Entrectinib | APPROVAL | Proto-oncogene tyrosine-protein kinase ROS inhibitor | 9.7 | IC50 | 33 |
| Crizotinib | APPROVAL | Proto-oncogene tyrosine-protein kinase ROS inhibitor | 9.22 | KI | 94 |
| Taletrectinib | PHASE_3 | Proto-oncogene tyrosine-protein kinase ROS inhibitor | — | — | 8 |

**Disease evidence.** Open Targets links ROS1 to 711 diseases. The top six rows are all cancers, and lung disease dominates.

| Disease | Association score | Evidence items |
|---|---|---|
| non-small cell lung carcinoma | 0.713 | 2640 |
| neoplasm | 0.615 | 1277 |
| lung adenocarcinoma | 0.591 | 552 |
| cancer | 0.543 | 502 |
| lung carcinoma | 0.433 | 746 |
| melanoma | 0.399 | 53 |

**Other layers.** CIViC holds 35 records for ROS1. The first rows describe the ROS1 fusion in Lung Non-small Cell Carcinoma with crizotinib at evidence level A, labelled Sensitivity/Response. A record for ROS1 G2032R together with the CD74::ROS1 fusion is labelled Resistance to crizotinib at level C. Another record pairs the same combination with cabozantinib at level D, labelled Sensitivity/Response.

In TCGA and GTEx (Toil recompute), 24 cancer types have both tumour and normal samples. In the six rows shown, the three kidney types sit well below normal. Kidney Papillary Cell Carcinoma has a mean tumour value of 0.57 against 5.19 in normal tissue (log2 fold -3.17). Cervical and Endocervical Cancer runs the other way, at 3.45 versus 0.21 (log2 fold 3.98), but with only 3 normal samples.

STRING v12 lists 1040 interaction partners. EML4 scores 925, the highest of the six shown, and KRAS scores 862.

## What it means and how to read it

Three facts are worth reading together.

**1. No dependency in the cell-line panel does not mean ROS1 is irrelevant.** The 0 of 1178 result, including 0 of 126 lung lines, says the knockout does not hurt the panel as a whole. A fusion-driven kinase would be expected to matter only in the few models that carry the fusion. The table reports lineages, not fusion status, so a small dependent subset could be hidden inside a lineage average. To test this, you would need the fusion calls for individual lines.

**2. The drug and evidence layers agree with each other.** Open Targets puts non-small cell lung carcinoma on top, with 2640 evidence items. CIViC's level A record also concerns lung carcinoma, and the drug list is dominated by approved ROS1 inhibitors. The most potent exact values are repotrectinib at pChEMBL 10.15 (IC50), lorlatinib at 10.0 (Ki), entrectinib at 9.7 (IC50) and crizotinib at 9.22 (Ki). Higher means stronger, but IC50 and Ki values come from different assay formats. Treat the ranking as a rough guide, not a head-to-head comparison.

**3. The trial counts need care.** They are matched by drug name: 94 for crizotinib, 51 for lorlatinib, 33 for entrectinib, 12 for repotrectinib and 8 for taletrectinib. Drugs with many indications will have trials unrelated to ROS1, so the counts are not ROS1 trial counts.

Expression is where the layers pull apart. Several tumour types show lower ROS1 than matched normal tissue, which is not what you would expect for an oncogene. Bulk RNA averages many samples and cannot separate a small fusion-positive group from the rest. Read the expression table as tissue context, not as evidence about the driver.

The interaction list also needs caution. EML4 is a plausible partner for a kinase fusion context. Some high-scoring partners, such as the receptor genes OR13G1 and TAS2R50, are surprising, and a combined STRING score does not say which evidence channel produced it. Check the channel before drawing biology from it.

## How to reproduce it in BioQ

1. Open BioQ and enter the gene symbol exactly as below.
2. Run the query and wait for the panels. In our run the engine took 0.59 seconds for 7 queries.
3. Read the dependency panel for the cell-line count and the share dependent, then the lineage rows.
4. Open the drug panel and note the exact-potency type (IC50 or Ki) next to each pChEMBL value.
5. Check the disease panel and the CIViC records, and note the evidence level and significance of each.

```text
ROS1
```

[Open this query in BioQ](https://bioq.bio/login?next=%2Fapp%23q%3DROS1)

Every panel names its source and release, so you can cite the DepMap release, the Open Targets Platform and CIViC separately. Your numbers may differ if the sources have updated since 2026-11-06.

## Limits: what the data do not show

- **Fusion status of cell lines.** The dependency table does not say which lines carry a ROS1 fusion, so a fusion-specific dependency cannot be ruled in or out.
- **Threshold choice.** A dependency is defined by a named cut-off on the Chronos effect. Another cut-off would give other counts.
- **Potency across assays.** pChEMBL values mix IC50 and Ki and come from different studies, so they do not rank drugs for any use.
- **Taletrectinib.** It is listed at PHASE_3 with no exact pChEMBL value in this view.
- **Clinical meaning.** CIViC levels describe evidence strength for a variant and therapy pairing. They are not guidance for treatment decisions.
- **Association scores.** Open Targets scores combine many evidence types. A high score does not show causation, and broad terms such as neoplasm and cancer overlap with the specific lung entries.
- **Expression.** Bulk tumour and normal comparisons do not isolate fusion-positive tumours, and some comparisons rest on very few normal samples.
- **Interaction scores.** STRING scores do not show direct binding.

## Data and method
The numbers on this page were read from the BioQ engine on 2026-11-06 (7 queries, 0.6 s of engine time in total). The tables are generated by code from the engine's answers; the text only restates them.

`/civic`, `/depmap`, `/drugs`, `/interaction`, `/ot_association`, `/tcga`

Run the same query yourself: [ROS1](https://bioq.bio/login?next=%2Fapp%23q%3DROS1)

*For research use only. This page summarises public data and is not medical advice.*

BioQ is an AI search engine and agent with one goal: making biology research 100X faster. Ask in the web app, over the REST API, or from Claude, ChatGPT or Cursor through MCP.

## Queries used

- [`ROS1`](https://bioq.bio/login?next=%2Fapp%23q%3DROS1)

## Frequently asked questions

### Is ROS1 a CRISPR dependency in DepMap?

Not in the DepMap release used here (depmap_24q4, Chronos). 0 of 1178 cell lines are dependent, the mean gene effect is 0.094, and ROS1 is not flagged as common essential. The panel does not report fusion status, so a dependency limited to fusion-positive models cannot be excluded.

### Which ROS1 inhibitors have the highest exact potency in ChEMBL?

Among the five drugs listed, repotrectinib has the highest exact value at pChEMBL 10.15 (IC50), followed by lorlatinib at 10.0 (Ki), entrectinib at 9.7 (IC50) and crizotinib at 9.22 (Ki). Because IC50 and Ki come from different assay types, the values are not a direct head-to-head comparison.

### Which disease is most associated with ROS1 in Open Targets?

Non-small cell lung carcinoma, with an association score of 0.713 and 2640 evidence items. ROS1 is linked to 711 diseases in total, and lung adenocarcinoma follows at 0.591.

### What does CIViC say about ROS1 resistance?

Of the 35 CIViC records for ROS1, one shown pairs ROS1 G2032R with the CD74::ROS1 fusion in Lung Adenocarcinoma. It is labelled Resistance to crizotinib at evidence level C. The records describe evidence, not treatment advice.

### How long does the combined ROS1 query take?

The engine ran 7 queries across 6 distinct resources in 0.59 seconds. That figure covers the engine only. Reading and checking the results takes additional time.

## Sources

- Dataset: [DepMap CRISPR Chronos gene effect (24Q4)](https://depmap.org/portal/)
- Dataset: [Open Targets Platform; ChEMBL; Drug Repurposing Hub](https://platform.opentargets.org/)
- Dataset: [TCGA pan-cancer expression (UCSC Xena Toil recompute)](https://xenabrowser.net/)
- Dataset: [CIViC: Clinical Interpretation of Variants in Cancer](https://civicdb.org/)
- Dataset: [STRING v12 protein-protein interactions](https://string-db.org/)

> Results come from public databases as of the date shown and may change. BioQ gives research information, not medical advice.
