Biology Atlas

Map / Open register

question · intermediate · partly written

Does assay metadata beyond a batch label improve prediction on held-out labs?

Batch labels are used and often subtracted; operator, site, date or quality as typed inputs have no measured gain on record.

What the sources show more

In an unbalanced reanalysis, adjusting for batch while keeping the group gave 2,011 genes against 11 when batch was blocked; IC50 spread was about 25% larger than Ki spread across public ChEMBL data. 12

Reported, not yet replicated

What is missing more

A multi-lab dataset with batch balanced against condition and richer fields recorded, scored on held-out labs against the batch label alone. 3

Our inference

What would settle it (our inference) more

Public multi-lab ADMET or CRISPR set: source lab, assay type and year as features against none and against the batch label alone, split by lab. 12

Our inference

Observation

Integration methods take one categorical batch label; richer provenance vocabularies exist but were not found used as model input. 34

Reported, not yet replicated

What is known

In an unbalanced reanalysis, adjusting for batch while keeping the group gave 2,011 genes against 11 when batch was blocked; IC50 spread was about 25% larger than Ki spread across public ChEMBL data. 12

Reported, not yet replicated

What is not

A multi-lab dataset with batch balanced against condition and richer fields recorded, scored on held-out labs against the batch label alone. 3

Our inference

What would settle it

Public multi-lab ADMET or CRISPR set: source lab, assay type and year as features against none and against the batch label alone, split by lab. 12

Our inference

Still open here

Not yet written: Hypotheses, What is missing, Who works on it.

Linked from

Sources

  1. Nygaard V, Rødland EA, Hovig E. Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses. Biostatistics (2016). paper; opened 2026-11-18 all records citing this
  2. Kalliokoski, T., et al.. Kalliokoski et al. 2013: Comparability of mixed IC50 data (PLoS ONE). PLoS ONE (2013). paper; opened 2026-11-17 all records citing this
  3. Luecken MD, et al.. Benchmarking atlas-level data integration in single-cell genomics. Nature Methods (2022). paper; opened 2026-11-18 all records citing this
  4. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics (2007). paper; opened 2026-11-18 all records citing this
Where it sits on the map
History of this record (5)

This record has been changed 5 times, on the dates below. The note kept against each change is written for the people who maintain the atlas, in their own shorthand, so it is not shown here; it is in the record itself.

Kind
amit
Register
open-question
Triage
needs-research
Status
partly written
Last updated
2026-11-19
Record
content/questions/does-richer-assay-metadata-beat-a-batch-label.json