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 H4K20me1_test.csv.gz against train set H4K20me1_train.csv.gz
Data Leakage:!WarningHow a flag is decided →
gb-qc v1.0.0 · run 2026-09-15 22:31:11

Basic Descriptive Statistics

Filename H4K20me1_test.csv.gz H4K20me1_train.csv.gz
Number of sequences 2270 30000
Minimum length 1000 1000
Mean length 1000.00 1000.00
Maximum length 1000 1000
!

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 H4K20me1_test.csv.gz H4K20me1_train.csv.gz
Data Leakage (percentage) 0.66% 0.09%
Data Leakage (count) 15 27
Similarity Histogram
29 high-similarity alignments
Query (Q) Target (T) Q Cov. T Cov. % Identity % Similarity E-value Alignment
seq_1266_test seq_5743_train 0.98 0.98 98.7 96.43 0.00e+00
seq_1266_test seq_15827_train 0.99 0.99 97.0 95.84 0.00e+00
seq_1811_test seq_18200_train 0.98 0.98 97.1 95.35 0.00e+00
seq_1347_test seq_26389_train 1.00 1.00 95.6 95.22 0.00e+00
seq_1362_test seq_6106_train 1.00 1.00 95.0 94.91 0.00e+00
seq_2084_test seq_14927_train 0.99 0.99 95.8 94.84 0.00e+00
seq_2252_test seq_15827_train 0.98 0.98 96.7 94.77 0.00e+00
seq_1811_test seq_4192_train 0.98 0.98 96.4 94.38 0.00e+00
seq_2263_test seq_28743_train 0.98 0.98 95.9 94.37 0.00e+00
seq_2099_test seq_12089_train 0.97 0.97 96.8 93.80 0.00e+00
seq_1924_test seq_12962_train 0.98 0.98 95.2 93.11 0.00e+00
seq_1362_test seq_17334_train 0.94 0.94 98.5 92.89 0.00e+00
seq_2084_test seq_13406_train 0.99 0.99 93.8 92.77 0.00e+00
seq_1362_test seq_29950_train 0.99 0.99 93.7 92.67 0.00e+00
seq_2099_test seq_7468_train 0.96 0.96 96.6 92.35 0.00e+00
seq_2252_test seq_5743_train 0.95 0.95 97.3 92.34 0.00e+00
seq_2084_test seq_2315_train 0.97 0.97 95.1 92.15 0.00e+00
seq_1191_test seq_16912_train 0.94 0.94 97.5 92.14 0.00e+00
seq_1413_test seq_25631_train 1.00 0.99 92.5 91.85 0.00e+00
seq_2224_test seq_1636_train 0.98 0.97 94.4 91.85 0.00e+00
seq_1442_test seq_23396_train 0.99 0.99 92.7 91.59 0.00e+00
seq_1543_test seq_15885_train 0.95 0.95 96.2 91.39 0.00e+00
seq_1362_test seq_13324_train 0.96 0.95 95.5 91.11 0.00e+00
seq_1854_test seq_19128_train 1.00 1.00 91.0 90.82 0.00e+00
seq_1362_test seq_24713_train 0.96 0.95 95.0 90.53 0.00e+00
seq_2263_test seq_23711_train 0.96 0.96 94.1 90.15 0.00e+00
seq_2224_test seq_21583_train 0.98 0.98 92.1 90.07 0.00e+00
seq_1924_test seq_15024_train 0.94 0.94 96.2 90.04 0.00e+00
seq_2084_test seq_15576_train 0.98 0.97 92.8 90.02 0.00e+00