NeoDiscoveryAdmin commited on
Commit
cbbc397
·
1 Parent(s): 83f28db

update instructions of v0.2

Browse files
Files changed (1) hide show
  1. README.md +13 -6
README.md CHANGED
@@ -102,17 +102,24 @@ These reflect the present scope of published evidence and curation coverage. Fut
102
 
103
  ## Versioning & Updates
104
 
105
- - **Current version: v0.1**
106
- This release contains 354 curated entries, covering 16 cancer types and 58 unique HLA alleles (MHC-I). All entries include experimental immunogenicity evidence. Peptide lengths range from 8–13 amino acids, with 9-mers being the most common.
 
 
107
 
108
  - **Update strategy**
109
- Future releases will expand coverage by adding new literature sources, additional mutation types, and molecular features not yet included in v0.1. Patient-level metadata will also be progressively enriched as more information becomes available.
 
 
 
110
 
111
  - **Changelog**
112
- All modifications and additions will be documented in `CHANGELOG.md`, including new data sources, field extensions, and corrections.
 
113
 
114
  - **Reproducibility**
115
- Historical versions of the dataset will remain available in the repository, allowing users to reproduce analyses and compare results across versions.
 
116
 
117
  ---
118
 
@@ -142,7 +149,7 @@ All contributions will be acknowledged in future version updates.
142
 
143
  If you use NDD in your research, please cite as:
144
 
145
- **Neoantigen Discovery Dataset (NDD), version 0.1**
146
  GitHub Repository: [https://github.com/NeoDiscovery/NDD](https://github.com/NeoDiscovery/NDD)
147
  Hugging Face Dataset: *to be released*
148
 
 
102
 
103
  ## Versioning & Updates
104
 
105
+ - **Current version: v0.2**
106
+ This release contains **257 curated positive neoantigen peptides**, covering **25 cancer types** and **46 unique HLA class I alleles**.
107
+ All entries are experimentally validated for immunogenicity and include harmonized molecular features such as binding rank, stability, anchor mutation, and expression context.
108
+ Peptide lengths range from **8–13 amino acids**, with **9-mers** remaining the most common.
109
 
110
  - **Update strategy**
111
+ NDD v0.2 represents a focused, high-quality subset of experimentally confirmed positive neoantigens,
112
+ refined to support transparent benchmarking and community collaboration.
113
+ Future versions will expand by incorporating validated negatives, additional mutation types,
114
+ and extended clinical metadata to enable comprehensive model training and evaluation.
115
 
116
  - **Changelog**
117
+ All modifications, new fields, and updates are documented in `CHANGELOG.md`,
118
+ including patient cohort adjustments and dataset schema consistency.
119
 
120
  - **Reproducibility**
121
+ Historical drafts prior to v0.2 have been deprecated and are no longer public,
122
+ ensuring that analyses and benchmarks reference a single, consistent dataset release.
123
 
124
  ---
125
 
 
149
 
150
  If you use NDD in your research, please cite as:
151
 
152
+ **Neoantigen Discovery Dataset (NDD), version 0.2**
153
  GitHub Repository: [https://github.com/NeoDiscovery/NDD](https://github.com/NeoDiscovery/NDD)
154
  Hugging Face Dataset: *to be released*
155