Use cases

Turn the time you spend gathering data into time spent asking

BioQ is not a tool for searching gene names one at a time. It is where a team asks a question and gets the evidence back. Here is who gets the most from it, when, and what to ask.

Who uses it, and when

The moment a research question comes up is the moment to use it. Every question below opens ready to ask.

Target discovery and validation teams

WhenBefore a candidate target goes to the team meeting, when you want the evidence on one page

Ask

Is KRAS a selective dependency in pancreatic cancer, and what has been measured binding it?

Layers read together
Cancer dependency (DepMap)binding measurements (ChEMBL, BindingDB)variants (ClinVar)expression (GTEx, HPA)structure (PDB, AlphaFold)
What you get

One answer brings a target’s dependency, drugs, variants and structure together, each with its source.

Ask this

Computation and AI teams

WhenWhen you want to give an in-house agent or an analysis pipeline tools for biological data

Ask

Use the BioQ tools to fetch the binding measurements for EGFR and gefitinib and put them in a table.

Layers read together
REST APIMCP (Claude, ChatGPT, Cursor)dozens of layers through the same tools
What you get

No downloading and cleaning first: the agent calls a tool and gets values that carry their evidence.

Ask this

Drug and medicinal chemistry teams

WhenWhen you are building a selectivity, off-target or repurposing hypothesis

Ask

Compare measured potencies of imatinib against ABL1 and KIT, and tell me how essential each is to cancer cells.

Layers read together
Raw binding measurementsdrug–targetcancer drug response (GDSC, PRISM)dependency
What you get

The questions that run from one drug to its targets to cell response stay in one conversation.

Ask this

Protein and biologics engineering teams

WhenWhen a new sequence arrives, or before you filter your designed candidates

Ask

What is the closest protein to this sequence, and which domains does it have? What does it look like folded?

Layers read together
Homology search (billions of sequences)domains (Pfam)protein language modelstructure neighbours (Foldseek)fold to 3Dimmunology (IEDB)
What you get

Paste a sequence and it runs on to the closest proteins, the domains and the structure. The language model never sees the sequence.

Ask this

Microbiome, enzyme and synthetic biology teams

WhenWhen you look for new enzyme candidates in a metagenome such as soil or gut

Ask

In the soil biome, which genes occur on the same contigs as protein clusters with the PF00069 domain?

Layers read together
Metagenomic proteinsbiomesco-occurrencegene neighbourhoodsPfam
What you get

Billions of metagenomic proteins are queried on the server, with nothing to download.

Ask this

Clinical and translational genetics teams

WhenWhen you sort out what a candidate variant means and the context around it (research use)

Ask

Which BRCA1 variants near chr17:43044345 are pathogenic, and in which tissues is the gene expressed, as which isoforms?

Layers read together
ClinVargene–disease (HPO, OMIM, ClinGen)GTEx isoformsUCSC link
What you get

Variants, disease and expression by tissue run together on one screen, and an uncertain variant is never promoted to pathogenic.

Ask this

Getting the most out of BioQ

The same tool answers very differently depending on how you ask. Eight habits worth learning first.

  1. Start with a question, not a name

    “Is KRAS essential in pancreatic cancer?” wakes more layers than “KRAS”. Put the disease, the cell line or the tissue in the same sentence.

  2. Narrow it with follow-ups

    The agent remembers the recent conversation. Ask “only the ones with a drug”, or “and this sequence?”, and carry on.

  3. Paste the sequence as it is

    Unpublished sequences are fine. The language model only sees a placeholder such as [SEQ1]; the server does the real search.

  4. Cross independent evidence with the overlay

    Where several layers point at the same organ or the same position is a lead worth checking. Read the evidence under the lead before you judge.

  5. Read “none” as four different things

    Searched and found nothing is not the same as never searched. Every answer says which one it is.

  6. Fold the sequence and look

    Paste a sequence and fold it: a 3D model in seconds. Do not trust the stretches below confidence (pLDDT) 70.

  7. For pipelines, the API and MCP

    The same tools work as a REST API and over MCP. One connector address puts them in Claude, ChatGPT or Cursor; copy results as a table and hand them to the next analysis.

  8. Carry the evidence into your report

    Every number comes with its source and identifier. Cite it as it is in a report or a meeting deck.

How BioQ differs from other tools

GeneCards and Open Targets are excellent and widely used in academia. BioQ is built for a different job. Here is which fits which task.

WhatBioQGeneCardsOpen Targets
Built forA search engine and agent that answers across genes, sequences, structures, drugs and metagenomes with one questionThe human gene database: one card per gene, with many sources brought togetherA platform that gathers evidence linking targets and diseases and scores it
How you askNatural-language conversation (it remembers follow-ups), a pasted sequence, identifiers and coordinatesSearch by gene name or keyword, then read the cardSearch a target or a disease, then read the scores and the evidence tables
Data scopeHuman genes plus billions of metagenomic proteins, tens of thousands of non-human genomes, domains, structures, immunology and raw binding measurementsCentred on human genes, with more than 190 sources brought togetherTarget–disease evidence: genetics, somatic mutations, drugs, pathways, expression, animal models, literature
Crossing layersAsks dozens of layers at once and points to where independent evidence agrees, marking a lead as a lead, not a discoveryLists a gene’s many sources on one cardAdds up scores per kind of evidence into a target–disease score
Where you startFrom a sequence, on to homologs, domains and structure; your own sequence is folded to 3D on the spotFrom a gene name or identifierFrom a target or a disease
AutomationOne REST API and one MCP server reach dozens of layers, for Claude, ChatGPT, Cursor or your own agentNot found in its public documentationA GraphQL API and its own MCP server (Open Targets data only)
TermsFree to try every month; company plans from $2,000 a monthFree for non-profit academic institutions and hospitals; companies need a licencePublic service with a public API

Compiled on 2026-11-04 from what each service publishes about itself. The cells summarise the main way each is used and do not list every feature. If something is not accurate, tell us: support@bioq.bio

When BioQ fits better

  • You want to go from one sequence to its closest proteins, domains and structure
  • You need to query metagenomes or non-human life at the scale of billions
  • You want to judge across layers at once and narrow down with follow-up questions
  • You connect it to an in-house agent or pipeline, or you handle confidential sequences

When another tool fits better

  • To skim one human gene on one page, GeneCards (more than 190 sources) is convenient.
  • To check the score and public evidence for a target–disease link, Open Targets is the reference. Its data can be queried in BioQ as well.
  • If you are a non-profit academic institution, GeneCards is free for you.

Bring BioQ to your discovery team

Start free with your own questions. When your team needs more, apply for a company plan; Enterprise adds private deployment and SSO.