The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 290, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to number in row 0
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 101, in _split_generators
pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 304, in _generate_tables
batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
examples = [ujson_loads(line) for line in original_batch.splitlines()]
~~~~~~~~~~~^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
return pd.io.json.ujson_loads(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
ValueError: Expected object or value
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 71, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
MitoSeqGen Dataset — Curated Vertebrate Mitochondrial Coding Sequences
This is the quality-controlled, deduplicated dataset used to train and evaluate MitoSeqGen, a constrained generative transformer for mammalian mitochondrial mRNA codon optimization. Training code: https://github.com/Maheshbonthada/MItoDNA
Source and construction
Coding sequences (CDS) for the 13 mitochondrial protein-coding genes (ND1, ND2, COX1, COX2, ATP8, ATP6, COX3, ND3, ND4L, ND4, ND5, ND6, CYTB) were bulk-downloaded from NCBI via the Entrez E-utilities across vertebrate mitochondrial genome records (244,874 candidate records). Each record passed an 8-layer QC pipeline (format integrity, nucleotide composition, CDS structural validity, genetic-code compliance under NCBI translation table 2, biological plausibility, exact-duplicate detection, taxonomic metadata completeness, dataset-level adequacy) — 100,250 records (40.9%) passed. Passing records were further deduplicated at the near-duplicate level (95% nucleotide identity, CD-HIT-style clustering), yielding 46,264 unique sequences spanning 4,134 species. Full methodology is in the training repo's manuscript.
This is not raw NCBI data redistribution — the raw genome files (GenBank format, ~700MB) are not included here since they are trivially re-obtainable directly from NCBI Entrez by anyone; this release is the curated, QC'd, deduplicated output specific to this project.
Files
| File | Description | Records |
|---|---|---|
mito_cds_deduped.fasta |
The curated dataset — final deduplicated CDS set used for train/val/test splitting | 46,264 |
mito_cds_passed.fasta |
All QC-passed records before near-duplicate clustering (shows provenance) | 100,250 |
mito_trna_passed.fasta |
QC-passed mitochondrial tRNA sequences (auxiliary; not used in the primary codon-optimization results) | — |
mito_cds_stats.json |
Summary statistics for the deduplicated set | — |
train.json, val.json, test.json |
Species-level (phylogenetic) train/val/test split indices — zero species overlap between splits | 36,471 / 4,638 / 5,155 |
train_augmented.json |
Training split with 2 additional RSCU-guided synonymous variants per sequence (the actual training data fed to the model) | 109,413 |
qc_report_2026-07-22.json |
Full QC audit trail: per-check pass/fail counts, rejection-reason breakdown, taxonomic/gene-length distributions | — |
Not included (regenerable from the training repo, or available upstream): raw NCBI GenBank downloads (re-obtainable from NCBI Entrez directly), and the tokenized/model-input-format JSON files (mito_cds_tokenized*.json, regenerable from mito_cds_deduped.fasta via src/data/preprocess.py in the training repo — these are a training-format serialization of the same sequences listed above, not distinct content).
Splitting methodology
Splits are by species identity, not by individual sequence, to prevent species-level data leakage (the same species' sequence appearing in both training and evaluation). This guarantees zero species overlap between train/val/test.
License
MIT — same as the training code and model weights.
Citation
Preprint in preparation (bioRxiv); citation details will be added here once posted. In the meantime, please cite via the GitHub repository: https://github.com/Maheshbonthada/MItoDNA
- Downloads last month
- -