A separate, higher-depth yeast reporter experiment measured a 71,103-sequence test library spanning random, native yeast, high/low-expression, challenging, SNV, motif-perturbation, and motif-tiling constructs. The library contains 80-bp variable inserts in a fixed promoter-like YFP reporter context; the processed table retains the source MAUDE expression/activity score and source pair annotations.
Processed tables are specific to each experiment. Column names, units, measurements, and table structure are not standardized across the database. Check this experiment’s column definitions and quality-control notes before comparing or combining data.
Basal / Untreated; grown in filtered Chardonnay grape must for 48 h
The authors cloned 80-bp random or designed DNA sequences into a promoter-like context upstream of yellow fluorescent protein (YFP), transformed the library into S288c delta-URA3 yeast, selected transformants in SD-Ura medium, grew the pool in Chardonnay grape must, sorted cells by fluorescence-activated cell sorting (FACS), and sequenced the sorted bins. The test library was assayed separately with approximately 100 cells sorted per sequence; MAUDE estimated the expression/activity score from read abundance across sorting bins. The full source sequence is 110 bp, consisting of the 80-bp variable region at full-sequence positions 17-96 (0-based inclusive) plus fixed reporter flanks.
Processed data
50 rows per page. Click a cell to inspect its full value.
Visible columns (14 of 14)
Row
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Page 1 · 50 rows · More results available
Filters apply to this table only. The CSV download contains the complete processed table; filtered rows are available through the API.
Column dictionary · 14 definitions
element_id
Stable package identifier for one test construct, derived from its zero-based source row position.
source_position_0based
Zero-based row position in the GEO/Zenodo filtered test data file and in the source subset position fields.
sequence
Complete 110-bp reporter construct sequence, including fixed flanks and the 80-bp variable insert.
variable_sequence_80bp
The 80-bp variable insert extracted from sequence positions 17-96 (0-based inclusive); fixed flanks are omitted.
sequence_length
Length of the complete reporter construct sequence in nucleotides; all retained rows are 110.
maude_expression
Source MAUDE-estimated expression/activity score for the construct; units and scaling are those supplied by the authors.
test_subset
Source test-subset membership, using semicolon-separated labels when a construct belongs to multiple subsets; unclassified_test marks rows not represented in the supplied subset archive.
source_tag
Source construct tag(s) from the subset annotation archive; multiple tags are semicolon-separated.
variant_type
Pairwise comparison type for source-linked pairs: SNV, motif_perturbation, or motif_tiling.
variant_role
Role within source-linked pairwise comparisons: alternate, reference, or both when the construct participates in multiple comparisons.
variant_pair_ids
Package-generated pair identifiers, with values aligned by semicolon to paired_element_ids, paired_expression_deltas, and variant_offsets_0based.
paired_element_ids
Package element IDs for source-linked comparison partners, aligned to variant_pair_ids.
paired_expression_deltas
Derived alternate-minus-reference MAUDE activity difference for each linked pair, aligned to variant_pair_ids; not a source-provided column.
variant_offsets_0based
Positions within the 80-bp variable insert at which the paired sequences differ, aligned to variant_pair_ids; multiple differing positions within one pair are pipe-separated and NA denotes no differing position in the variable insert.
Quality control
The source GEO file is already described as filtered and contains 71,103 MAUDE-scored test constructs. Package-level QC required exactly two tab-separated fields, a finite numeric activity value, exactly 110 nucleotides, only uppercase A/C/G/T bases, and the expected fixed reporter flanks; all 71,103 source rows passed and no rows were removed. Every position referenced by the source subset archive was checked against the corresponding sequence, and all paired/single sequence annotations matched. Exact duplicate full sequences were absent. The source pair relationships were retained even when a construct participates in multiple comparisons; 8,911 rows not assigned to a named source subset are retained as unclassified_test rather than discarded.
Curation notes
This package represents the authors' separate, higher-depth held-out test MPRA experiment rather than the 6,739,258-row training library. The source test file contains 71,103 measured constructs; it is not a prediction table. Subset annotations in the supplied archive overlap because some random/native constructs are reused in variant trajectories and some constructs participate in multiple pairwise tests. The source archive contains 46,236 SNV pair records, 3,527 motif-perturbation pair records, and 2,653 motif-tiling pair records; these counts differ from rounded/older counts reported in the paper's summary table, so the source archive is preserved as authoritative for the row-level annotations. The table's paired_expression_deltas are computed from the source MAUDE scores to make variant effects directly queryable. No genomic assembly is assigned because the library mixes synthetic constructs with native yeast promoter fragments and the source does not specify an assembly for this test table.