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_detection_300_tata_test.csv.gz against train set promoter_detection_300_tata_train.csv.gz
Data Leakage:✖FailHow a flag is decided →
gb-qc v1.0.0 · run 2026-09-15 22:30:51

Basic Descriptive Statistics

Filename promoter_detection_300_tata_test.csv.gz promoter_detection_300_tata_train.csv.gz
Number of sequences 613 4904
Minimum length 300 300
Mean length 300.00 300.00
Maximum length 300 300
✖

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_detection_300_tata_test.csv.gz promoter_detection_300_tata_train.csv.gz
Data Leakage (percentage) 3.43% 0.47%
Data Leakage (count) 21 23
Similarity Histogram
24 high-similarity alignments
Query (Q) Target (T) Q Cov. T Cov. % Identity % Similarity E-value Alignment
seq_452_test seq_4014_train 1.00 1.00 100.0 100.00 4.30e-165
seq_568_test seq_2758_train 1.00 1.00 100.0 100.00 4.30e-165
seq_88_test seq_1777_train 1.00 1.00 100.0 100.00 4.30e-165
seq_63_test seq_556_train 1.00 1.00 100.0 100.00 4.30e-165
seq_5_test seq_4839_train 1.00 1.00 100.0 100.00 4.30e-165
seq_195_test seq_2345_train 1.00 1.00 100.0 100.00 4.30e-165
seq_272_test seq_3559_train 1.00 1.00 100.0 99.70 2.04e-164
seq_476_test seq_786_train 0.99 0.99 100.0 99.30 9.66e-164
seq_231_test seq_2028_train 0.99 0.99 100.0 99.30 9.66e-164
seq_210_test seq_1955_train 0.99 0.99 100.0 99.30 9.66e-164
seq_534_test seq_1109_train 0.99 0.99 100.0 99.30 9.66e-164
seq_182_test seq_2934_train 0.99 0.99 100.0 99.30 9.66e-164
seq_561_test seq_2939_train 1.00 0.99 99.6 98.90 2.20e-161
seq_148_test seq_1247_train 1.00 1.00 99.0 98.70 1.07e-158
seq_89_test seq_3636_train 1.00 1.00 98.6 98.60 2.31e-158
seq_70_test seq_4638_train 1.00 0.99 98.6 97.32 4.98e-158
seq_452_test seq_1183_train 1.00 1.00 96.6 96.60 2.10e-148
seq_184_test seq_389_train 0.99 0.99 96.3 95.63 1.88e-145
seq_229_test seq_2683_train 0.99 1.00 95.3 94.35 3.22e-141
seq_558_test seq_3658_train 0.99 0.99 94.9 93.95 2.62e-138
seq_550_test seq_389_train 1.00 1.00 92.6 92.32 8.24e-128
seq_558_test seq_4457_train 0.99 0.99 92.9 91.97 9.40e-129
seq_558_test seq_3684_train 0.99 0.99 92.2 91.00 5.37e-125
seq_267_test seq_534_train 1.00 0.99 91.7 90.51 5.37e-125