Discover 
HLA-Compass

Web-accessible HLA peptide database

subpage_hero_icon Alithea Bio
hla-compass

Alithea’s web-accessible HLA peptide database HLA Compass comprises 1.4 million unique peptides and 10,000 cancer specific epitopes deriving from >17 million identifications in 4,160 high-quality HLA peptidomics samples, meticulously collected from healthy organs, cancer cell lines, and tumor tissues. Each peptide in this extensive database is grounded in real-world  biological data, reflecting the actual peptide presentation on HLA complexes in these diverse tissue types.

Get the HLA-Compass

Unique benefits of HLA-Compass

bionformatics_pipeline Alithea

Unified bioinformatics pipeline based on Alithea’s deep domain knowledge and bioinformatic tools

All peptides are linked to quantitative information allowing for relative quantification between thousands of samples

Intuitive user interface allows for easy browsing of peptide and protein data

Tools to identify off-targets from lists of potential candidates or similarity searching (e.g. with x-scan data)

HLA_alleles Alithea

All peptides are mapped to cognate HLA alleles through Alithea’s proprietary HLA binding predictors

Stringent quality control based on Alithea’s deep domain knowledge and bioinformatic tools

The numbers behind HLA-compass

Alithea’s web-accessible HLA peptide database HLA Compass comprises 1.4 million unique peptides and 10,000 cancer specific epitopes deriving from >17 million identifications in 4,160 high-quality HLA peptidomics samples, meticulously collected from healthy organs, cancer cell lines, and tumor tissues.

Each peptide in this extensive database is grounded in real-world biological data, reflecting the actual peptide presentation on HLA complexes in these diverse tissue types.

Contact us for more details
solutions_image Alithea Bio

HLA Compass access options 
on “per-request” basis:

  • interrogating presentation of single peptides via peptide-centric analysis
  • interrogating presentation of all peptides from certain genes via gene-centric analysis
  • getting evidence for the presentation of putative off-targets y submitting lists of peptides for off-target analysis
  • identifying potential off-targets through off-target analysis (link) based on Alanine or x-scan data
  • re-searching subsets of the data for evidence that putative n

HLA-Compass

AI that understands immunopeptidomics.

Ask a question. The copilot queries proprietary HLA data, picks the right MCP tool, and runs a governed analysis — so you can discover neoantigens in minutes, not months.

How it works

From question to governed answer in three steps

Step 1

Ask in natural language

Describe your research question — the copilot understands immunopeptidomics terminology, your data context, and available tools.

Step 2

AI selects and chains tools

The copilot reads MCP descriptors from the tool marketplace, selects the right analytical module, configures inputs, and plans the execution.

Step 3

Governed execution with audit trail

Every action is logged, budget-controlled, and traceable. You get results with full provenance — which data, which tool, which model.

AI Copilot Playground

Ask a question. Get a governed answer.

Interact with our autonomous immunopeptidomics assistant. Select a prompt preset below, or type a custom question to see how the AI Copilot maps datasets and executes governed MCP tools in real-time.

HLA-Compass AI Copilot
PORT 9002 · ACTIVE
Which peptides from KRAS G12D are presented on HLA-A*02:01 in our lung tissue dataset?

I will query the local ligandome database for KRAS G12D neoepitopes restricted to HLA-A*02:01, then filter against your active lung tissue cohorts.

MCP TOOL · LIGANDOME-QUERY
source: lung_tissue_batch_04 · allele: A*02:01 · gene: KRAS · mutation: G12D
● 3 peptides found · 2 above presentation threshold

Found 3 peptides from KRAS G12D on HLA-A*02:01 in lung_tissue_batch_04. Two pass the presentation threshold with high confidence (q-value < 0.01).

Governed Audit #LQ-2847 0.3 ACT

How HLA-Compass compares with alternatives

Different approaches win for different jobs. Public atlases are better for open, citable reference checks, prediction tools are faster for sequence ranking, specialist services are stronger for bespoke wet-lab programs, and in-house stacks maximize control at the cost of time.

HLA-Compass Public atlasesHLA Ligand Atlas / SysteMHC Atlas Prediction toolsNetMHCpan / MHCflurry Specialist servicesPure MHC / CRO workflows In-house stack
Experimentally observed HLA peptides If instrumented
Open/public reference access
Private workspace for proprietary programs Project-based
Healthy tissue context for off-target review Partial Partial Partial
Self-serve interactive usage Partial
Governed execution / audit layer Possible
Custom workflow control Partial Partial
Time to first result Immediate Immediate Immediate Weeks Months