Ncrna Pretrain 47M: Non-Coding RNA Pretraining Corpus

syjiang's Non-coding RNA Pretraining Corpus contains 47,154,121 unique ncRNA sequences, representing 32.48 billion nucleotides. The dataset was aggregated from five complementary repositories, deduplicated, and clustered for balanced sampling. It was last updated on 2026-06-22.

Use Cases

• Pretrain RNA foundation models based on the large corpus of 47 million sequences.
• Benchmark sequence clustering algorithms based on the provided MMseqs2 linclust clusters.
• Study non-coding RNA diversity and distribution based on the aggregated, deduplicated sequences.

Strengths

• Contains 47,154,121 unique ncRNA sequences.
• Aggregates data from five complementary repositories, suggesting broad source coverage.
• Sequences are exactly deduplicated and clustered for balanced sampling.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified as the last update is dated 2026-06-22.

Provenance

Source

Aggregated from five complementary repositories (unspecified).

Collection Method

Sequences were unified to the RNA alphabet, exactly deduplicated, and clustered.

Freshness

Last updated 2026-06-22 03:03:00