# ProteoBench Astral DIA Entrapment — OpenDIA benchmark bundle ## Overview This package contains source data and OpenDIA-derived files for testing library-free DIA analysis on the **ProteoBench DIA Ion Entrapment – Astral** benchmark. The benchmark is designed to assess false-discovery-rate (FDR) control using matched target and entrapment peptide sequences. The source dataset contains three technical replicates of a **50 ng HeLa digest** acquired on a **Thermo Fisher Orbitrap Astral** in DIA mode with a **15-minute gradient**. The ProteoBench benchmark documentation describes: - 300 × 2 Th DIA isolation windows spanning 380–980 m/z; - HCD fragmentation at normalized collision energy 25; - Astral MS2 detection; - a pre-digested peptide FASTA containing human target peptides, contaminants, and matched entrapment peptides. Benchmark documentation: https://proteobench.readthedocs.io/en/v0.18.1/available-modules/active-modules/13-entrapment-ion-dia-astral/ The FASTA is **pre-digested**: each FASTA entry represents a peptide rather than a full-length protein. **Do not apply enzymatic digestion to this FASTA when reproducing the benchmark.** --- ## Package layout ```text . ├── README.md ├── MANIFEST.tsv ├── 100K_sampled_fasta │ ├── opendia_results │ │ ├── LFQ_Astral_DIA_15min_50ng_Human_01_extracted.chrom.xic │ │ ├── LFQ_Astral_DIA_15min_50ng_Human_02_extracted.chrom.xic │ │ ├── LFQ_Astral_DIA_15min_50ng_Human_03_extracted.chrom.xic │ │ ├── logs │ │ │ ├── opendia.stderr.log │ │ │ ├── opendia.stdout.log │ │ │ └── time.txt │ │ ├── meta │ │ │ ├── command.sh │ │ │ ├── exit_code.txt │ │ │ ├── fasta.sha256 │ │ │ ├── opendia.helphelp.txt │ │ │ ├── raw_files.txt │ │ │ ├── results_manifest.tsv │ │ │ ├── run_manifest.txt │ │ │ └── sif.sha256 │ │ └── OpenDIA.results.tsv │ ├── ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment_pep_100k_randomly_sampled.decoys.tsv.zst │ ├── ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment_pep_100k_randomly_sampled.decoys.tsv.zst.sha256 │ └── ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment_pep_100k_randomly_sampled.fasta ├── full_fasta │ ├── ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment.decoys.tsv.zst │ ├── ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment.decoys.tsv.zst.sha256 │ └── ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment_pep.fasta └── raw ├── LFQ_Astral_DIA_15min_50ng_Human_01.raw ├── LFQ_Astral_DIA_15min_50ng_Human_02.raw └── LFQ_Astral_DIA_15min_50ng_Human_03.raw ``` A machine-readable description of every file is provided in `MANIFEST.tsv`. --- ## Source data ### RAW files The three files under `raw/` are the ProteoBench Astral DIA entrapment benchmark technical replicates: ```text LFQ_Astral_DIA_15min_50ng_Human_01.raw LFQ_Astral_DIA_15min_50ng_Human_02.raw LFQ_Astral_DIA_15min_50ng_Human_03.raw ``` ProteoBench requests that these RAW filenames are **not renamed**, because run mapping in the benchmark relies on the original filenames. ### Full benchmark FASTA ```text full_fasta/ ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment_pep.fasta ``` This is the full pre-digested ProteoBench entrapment search space used for the full-library experiment. It contains peptide entries rather than intact proteins and includes human target peptides, contaminants, and matched entrapment sequences supplied by the benchmark. Because this is a peptide FASTA, conventional protein-level interpretation and protein mapping should be treated cautiously. ### 100K sampled FASTA ```text 100K_sampled_fasta/ ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment_pep_100k_randomly_sampled.fasta ``` This is a randomly sampled **100,000-entry subset** of the pre-digested benchmark FASTA. It was used as a smaller development and validation search space for OpenDIA. The exact OpenDIA command, input FASTA hash, container hash, input RAW list, and run metadata for this analysis are retained under: ```text 100K_sampled_fasta/opendia_results/meta/ ``` Those files should be treated as the authoritative reproducibility record for the 100K OpenDIA run. --- ## OpenDIA analysis The directory: ```text 100K_sampled_fasta/opendia_results/ ``` contains the completed OpenDIA analysis of the three Astral RAW files against the 100K sampled peptide FASTA. The analysis used OpenDIA's native FASTA-driven, library-free workflow: ```text pre-digested peptide FASTA ↓ peptidoform / precursor enumeration ↓ PeptDeep / ONNX assay prediction ↓ target + OpenDIA-generated decoy assay library ↓ DIA chromatogram extraction ↓ OpenSWATH scoring ↓ statistical rescoring / FDR control ↓ OpenDIA results ``` Important characteristics of the benchmark/search configuration used in this project include: - no enzymatic cleavage, because the FASTA is already pre-digested; - precursor charge range 1–5; - precursor m/z range 400–1000; - fragment/product m/z range 100–1800; - fixed carbamidomethylation of cysteine; - methionine oxidation as the supported variable modification; - Astral/ThermoTOF PeptDeep prediction profile; - normalized collision energy 25 for prediction; - ion-mobility prediction/scoring disabled for this Astral dataset; - OpenDIA-generated target/decoy assay library. For the **exact** command and settings used for the packaged 100K run, use the captured metadata rather than this summary: ```text opendia_results/meta/command.sh opendia_results/meta/run_manifest.txt opendia_results/meta/opendia.helphelp.txt opendia_results/meta/sif.sha256 opendia_results/meta/fasta.sha256 opendia_results/meta/raw_files.txt opendia_results/logs/opendia.stdout.log opendia_results/logs/opendia.stderr.log opendia_results/meta/exit_code.txt ``` If a value in this README differs from the captured run metadata, the captured metadata is authoritative. --- ## Predicted target/decoy assay libraries Two compressed TSV assay libraries are included. ### 100K search-space library ```text 100K_sampled_fasta/ ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment_pep_100k_randomly_sampled.decoys.tsv.zst ``` This library was generated from the 100K sampled peptide FASTA and contains the predicted OpenDIA/OpenMS assay search space, including generated decoys. ### Full search-space library ```text full_fasta/ ProteoBenchFASTA_Entrapment_Human_with_contaminants_entrapment.decoys.tsv.zst ``` This is the corresponding predicted target/decoy assay library generated from the full pre-digested ProteoBench FASTA. These libraries are **derived files**; they are not original ProteoBench source files. Both TSVs are compressed using Zstandard and have adjacent SHA-256 checksum files. Verify a compressed file with: ```bash sha256sum -c .zst.sha256 ``` Decompress while retaining the compressed copy with: ```bash zstd -d -k .zst ``` --- ## OpenDIA result files ### `OpenDIA.results.tsv` ```text 100K_sampled_fasta/opendia_results/OpenDIA.results.tsv ``` This is the primary tabular OpenDIA result for the 100K sampled-FASTA analysis. Use `meta/command.sh` and `meta/run_manifest.txt` to determine the exact filtering, inference, and export settings associated with this result. ### Extracted chromatograms The files: ```text LFQ_Astral_DIA_15min_50ng_Human_01_extracted.chrom.xic LFQ_Astral_DIA_15min_50ng_Human_02_extracted.chrom.xic LFQ_Astral_DIA_15min_50ng_Human_03_extracted.chrom.xic ``` are OpenDIA-generated extracted-chromatogram artifacts corresponding to the three RAW runs. They are intermediate/reuse artifacts rather than the primary tabular identification result. ### Logs ```text opendia_results/logs/opendia.stdout.log opendia_results/logs/opendia.stderr.log opendia_results/logs/time.txt ``` These record program output, warnings/errors, and runtime information. ### Reproducibility metadata ```text opendia_results/meta/command.sh opendia_results/meta/exit_code.txt opendia_results/meta/fasta.sha256 opendia_results/meta/opendia.helphelp.txt opendia_results/meta/raw_files.txt opendia_results/meta/results_manifest.tsv opendia_results/meta/run_manifest.txt opendia_results/meta/sif.sha256 ``` These files capture the exact analysis invocation and software/input identities. --- ## Relationship between the two FASTA directories | Directory | Search space | OpenDIA results included | Predicted target/decoy library | |---|---|---:|---:| | `100K_sampled_fasta/` | 100,000-entry sampled peptide FASTA | Yes | Yes | | `full_fasta/` | Full ProteoBench peptide FASTA | No result table in this package | Yes | The 100K subset was used as a reduced development/validation search space. The full FASTA represents the complete benchmark search space. Identification counts from these two search spaces should **not** be compared as though the searches were otherwise identical: changing search-space size changes target/decoy competition and statistical calibration. --- ## Entrapment benchmark interpretation The ProteoBench entrapment benchmark evaluates whether a reported FDR is empirically calibrated. Entrapment peptides are designed such that they should not be present in the biological sample. Identifications assigned to those sequences therefore provide an empirical estimate of false discoveries. The current ProteoBench Astral entrapment module evaluates FDR at the **global precursor-ion level**. For current FDP/FDR definitions and benchmark interpretation, refer to ProteoBench: https://proteobench.readthedocs.io/en/v0.18.1/available-modules/active-modules/13-entrapment-ion-dia-astral/ --- ## Important caveats 1. **The ProteoBench FASTA is pre-digested.** Do not apply enzymatic digestion. 2. **Do not rename the RAW files** when using the ProteoBench benchmark workflow. 3. **Protein inference should be interpreted cautiously.** The benchmark search space is peptide-based rather than a conventional intact-protein FASTA. 4. `100K_sampled_fasta/` is a reduced development/validation search space, not a substitute for the complete benchmark search space. 5. The `*.decoys.tsv.zst` files are OpenDIA/OpenMS-derived assay libraries, not original ProteoBench source files. 6. Exact OpenDIA parameters for the included 100K result are recorded under `opendia_results/meta/`. --- ## Integrity checking The two compressed assay libraries have adjacent SHA-256 checksum files. For a complete checksum manifest of the whole shared directory, run from the package root: ```bash find . -type f \ ! -name 'MANIFEST.sha256' \ -print0 \ | sort -z \ | xargs -0 sha256sum \ > MANIFEST.sha256 ``` Verify the package later with: ```bash sha256sum -c MANIFEST.sha256 ``` --- ## Suggested acknowledgement When using the source benchmark data, cite or acknowledge **ProteoBench** and the **DIA Ion Entrapment – Astral** benchmark according to the current ProteoBench guidance. When using the derived assay libraries or OpenDIA results, also report the relevant **OpenMS/OpenDIA software version or commit** recorded in the accompanying run metadata and container checksum. --- ## Provenance note This bundle was assembled as part of OpenMS/OpenDIA library-free DIA scalability and FDR-validation work. It deliberately separates: ```text ProteoBench source inputs from OpenDIA/OpenMS-derived search libraries and results ``` Original source-data ownership and licensing remain with their respective providers.