Summary

Basic Descriptive Statistics
! Data Leakage
Documentation GitHub
Find the shortcut. Learn the biology.
Automated quality control for genomic machine learning datasets: scores the biases, duplicates and data leakage a classifier could exploit before you train on it.
Searching test set promoter_strength_protoplast_test.csv.gz against train set promoter_strength_protoplast_train.csv.gz
Data Leakage:!WarningHow a flag is decided →
gb-qc v1.0.0 · run 2026-09-15 22:31:29

Basic Descriptive Statistics

Filename promoter_strength_protoplast_test.csv.gz promoter_strength_protoplast_train.csv.gz
Number of sequences 1000 8000
Minimum length 170 170
Mean length 170.00 170.00
Maximum length 170 170
!

Data Leakage

Every test sequence searched against the train set with MMseqs2, and scored on its best hit. Similarity is min(query coverage, target coverage) × percent identity: how much of the shorter sequence the alignment spans, times how much of that matches. The histogram is the distribution of those best hits.

The flag is the percentage of test sequences at or above the threshold: Pass at 0%, Warning below 2%, Fail at 2% or more. A test sequence that near-duplicates a training one is one the model has already seen, so the accuracy it earns there is not evidence. What to do about it →

The panel under the histogram lists those alignments, up to the first 100; every one of them, listed or not, is exported to mmseqs/mmseqs2_search_result.tsv beside this report. Its columns are:

Query (Q)
Identifier of the test sequence: seq_<i>_test, where i is its 0-based position in the input test file.
Target (T)
Identifier of the train sequence: seq_<i>_train, where i is its 0-based position in the input train file.
Q Cov.
Fraction of the query sequence covered by the alignment (0–1).
T Cov.
Fraction of the target sequence covered by the alignment (0–1).
% Identity
Percent identical aligned positions in the aligned region.
% Similarity
Similarity score used for leakage detection, calculated as min(Q Cov., T Cov.) × % Identity.
E-value
MMseqs2 E-value (lower is more "significant").
Alignment
Click “Show” to expand the alignment visualisation for a row.
Filename promoter_strength_protoplast_test.csv.gz promoter_strength_protoplast_train.csv.gz
Data Leakage (percentage) 0.70% 0.29%
Data Leakage (count) 7 23
Similarity Histogram
23 high-similarity alignments
Query (Q) Target (T) Q Cov. T Cov. % Identity % Similarity E-value Alignment
seq_742_test seq_6675_train 1.00 1.00 98.8 98.80 3.57e-88
seq_665_test seq_6364_train 1.00 1.00 98.8 98.80 3.57e-88
seq_742_test seq_5047_train 1.00 0.99 98.2 97.61 3.78e-85
seq_633_test seq_647_train 1.00 1.00 97.6 97.60 8.17e-85
seq_224_test seq_2084_train 1.00 1.00 97.0 97.00 3.80e-83
seq_665_test seq_3372_train 1.00 1.00 97.0 97.00 3.80e-83
seq_660_test seq_166_train 1.00 1.00 96.4 96.40 1.74e-81
seq_633_test seq_6533_train 0.99 1.00 96.4 95.82 3.64e-80
seq_633_test seq_3064_train 1.00 1.00 95.8 95.80 7.78e-80
seq_602_test seq_5964_train 0.98 0.99 97.0 95.25 1.74e-81
seq_633_test seq_1766_train 1.00 1.00 95.2 95.20 3.42e-78
seq_633_test seq_7014_train 1.00 1.00 95.2 95.20 3.42e-78
seq_660_test seq_4522_train 1.00 1.00 94.7 94.70 1.47e-76
seq_243_test seq_298_train 0.98 0.99 97.0 94.67 7.97e-81
seq_633_test seq_4200_train 0.99 1.00 94.1 93.54 1.21e-73
seq_633_test seq_1415_train 1.00 0.98 95.2 93.49 6.57e-76
seq_633_test seq_5514_train 0.99 1.00 94.1 92.97 5.75e-74
seq_633_test seq_5589_train 1.00 0.99 94.1 92.97 5.75e-74
seq_633_test seq_592_train 0.99 0.99 93.4 92.84 1.11e-72
seq_633_test seq_1008_train 0.99 0.98 94.0 92.31 1.02e-71
seq_633_test seq_2524_train 0.99 0.99 93.4 92.28 2.12e-71
seq_660_test seq_3547_train 1.00 0.99 92.3 91.75 5.68e-67
seq_633_test seq_4687_train 0.99 0.99 92.3 91.19 3.15e-68