Dataset

nucleotide_benchmarks:splice_sites_acceptors

nucleotide_benchmarks:splice_sites_acceptors

Not in the Core Set 10 of the 93 cleared all five gates

Task: Binary classification, RNA Processing Prediction
Level: RNA
Species: Homo sapiens
Sequences: 33,000 across 2 splits
Sequence length: 600 to 600
Licence: CC BY-NC-SA 4.0

Read the 3 full reports for this dataset Every distribution, probe fit and overlapping sequence behind the numbers below

Warning worst flag across the ten checks on train, this dataset's reference split
0 of the ten checks returned Fail on that split 4 returned Warning, 0 could not be scored
0.699 AUROC from the strongest single sequence feature, which is Per position nucleotide content A lower bound on what a model could recover from the sequences alone
0.60% of one split's sequences were also present in another, at worst The largest of 1 split comparison: train vs test
0.995 AUROC when all those features are read together rather than one at a time Measured on a held-out split. This is what Core Set membership turns on

Where this sits in the corpus

Stronger than 58 of the 93 probed datasets
Higher than 90 of the 92 jointly scored datasets
More overlap than 38 of the 92 scanned datasets

A rank among the datasets this run measured, not among benchmarks in general.

What the checks returned

Worst flag across the ten checks on the train split: Warning

Of the nine checks that measure a single split, 9 ran on train and 9 were scored: 0 returned Fail, 4 returned Warning, 0 returned Unknown. The leakage scan is counted in the flag above but not in these figures, because it compares pairs of splits rather than measuring one.

Every rollup figure on this page was measured on train, which is this dataset's reference split — not a maximum over its splits.

The strongest single feature is Per position nucleotide content, which reaches AUROC 0.699 on the train split from that feature alone. That is a lower bound: a stronger feature, or a model, may recover more.

The leakage scan found that 0.60% of the sequences in test were also present in train; the same comparison the other way puts 0.12% of train in test. That is the largest of the one split comparison the scan made here.

What each of the ten checks measures · What a flag is, and what it does not say · Why one split is quoted

The reports behind those flags

The full report genomic-benchmarks-qc rendered for each split and each split pair, with every distribution, every probe fit, and every individually-listed overlapping sequence behind the numbers on this page. Each opens inside this site, and the file the tool wrote is one click further on.

train · sequence class distributions
test · sequence class distributions
train vs test leakage scan

Worked examples of a report

The same features, read together

Every flag above is one feature on its own. This is all of them at once: length, GC, mononucleotide and dinucleotide content and the two per-position profiles, handed to a single model and scored on a held-out split. Nothing is added that the checks do not already look at — the only change is that they are read jointly.

0.995 AUROC from all the checks’ features at once, measured on the test split 95% interval 0.993 to 0.996, over 3,000 held-out sequences

The strongest single feature reaches 0.699 on its own.

How it was measured, why it exists, and the six datasets it changes the reading of

The nine per-split checks

One row per split, one column per check. The leakage scan is not here because it compares split pairs rather than measuring one split; it has its own section below.

The nine class-distribution checks on each of nucleotide_benchmarks:splice_sites_acceptors’s 2 splits. The rollup reports the train row.

Table view

Flag and, where measured, the strongest probe AUROC per check per split.
Split	GC content	Nucleotide content	Dinucleotide content	Sequence lengths	Per position nucleotide	Per position from 3′	Duplicates within label	Duplicates across labels	Unique bases
train	Pass 0.584	Pass 0.585	Warning 0.628	Pass 0.500	Warning 0.699	Warning 0.699	Warning	Pass	Pass
test	Warning 0.636	Warning 0.628	Warning 0.649	Pass 0.500	Fail 0.704	Fail 0.704	Warning	Pass	Pass

Each split, the flag it rolled up to, its strongest probe, and the report every flag on that row was read out of.
Split	Sequence columns	Worst flag	Checks scored	Strongest probe	AUROC	Report
`train` reference	`sequence`	Warning	9	Per position nucleotide content	0.699	sequence
`test`	`sequence`	Fail	9	Per position nucleotide content	0.704	sequence

How far each probe gets

The AUROC each of the six probe checks reaches on each split, from that one feature alone. The axis starts at 0.50 because that is what a feature carrying no class information returns.

What each single-feature probe recovered, on each split. Where a dataset has several sequence columns the strongest is shown; the table view lists every value.

The axis starts at 0.50 because that is what a probe recovering nothing would read. A value is a lower bound: it is what one feature reached, not the most that could be reached.

Table view

Probe AUROC per check per split.
Check	train	test
Per sequence GC content	0.584	0.636
Per sequence nucleotide content	0.585	0.628
Per sequence dinucleotide content	0.628	0.649
Sequence lengths	0.500	0.500
Per position nucleotide content	0.699	0.704
Per position reversed nucleotide content	0.699	0.704

Split comparisons

For every pair of splits the scan compared, the share of each split whose sequences also appear in the other. Each bar is coloured by the split it counts.

The two directions of a pair are different statements. A held-out split largely contained in train is the one that bears on whether a score on it is a held-out score.

Table view

Exact-sequence overlap, both directions, per split pair.
Comparison	Share	Flag
train also in test	0.12%	Warning
test also in train	0.60%	Warning

Each pair, in both directions. The two percentages differ whenever the two splits differ in size, so both are shown rather than the larger one.
Comparison	Flag	Of the first split	Of the second
`test` in `train`	Warning	0.60%	0.12%

Licence, credit, and the data itself

What this data may be used for is its source's decision and not this project's. It is set down here as the curation read it, with the evidence, so that a reader can check the reading rather than take it.

Licence: CC BY-NC-SA 4.0
Credit: 10.1038/s41592-024-02523-z — the collection's

Each split's file, and where it is served from:

train — train.parquethuggingface.co
test — test.parquethuggingface.co

This project may not republish these files. What is linked above is the source's own copy and the only one there is — nothing here mirrors it, and a reader who wants the data has to fetch it from them. The dataset card states no licence; the Nucleotide Transformer project repository is CC BY-NC-SA 4.0. InstaDeep were asked and declined permission to re-upload (September 2026). These 32 splits are not republished; the files stay at InstaDeepAI/nucleotide_transformer_downstream_tasks_revised, which is what dataset_url names for them.

How to cite the run these numbers come from

Nearby

Nearest by name in nucleotide_benchmarks: nucleotide_benchmarks:H3K9ac, nucleotide_benchmarks:H3K9me3, nucleotide_benchmarks:H4K20me1, nucleotide_benchmarks:enhancers, nucleotide_benchmarks:promoter_all, nucleotide_benchmarks:promoter_no_tata, nucleotide_benchmarks:promoter_tata, nucleotide_benchmarks:splice_sites_donors · and 7 more