One culture. One connected story.

The same cell-line name does not guarantee the same model

Ben-David and colleagues analysed 106 human cancer cell lines grown in two laboratories and characterised 27 strains of MCF7 in more depth. Testing those 27 strains against 321 anticancer compounds revealed substantial differences: at least 75% of compounds that strongly inhibited some strains were inactive in others. That percentage refers to the relevant active compounds in this experiment, not to all drugs or all cancer research. ^[1]

The Broad Institute’s account of the work explained why evolution during culture can contribute to disagreement between studies using nominally the same line. This is reporting on the same study, not an independent replication. Together with the primary paper, it provides a clear reason to preserve the history of a particular culture rather than relying on its name alone. ^[2]

Documentation cannot detect a genomic change by itself. The recommendation here is to retain the context needed to interpret characterisation data and investigate differences. A complete history complements appropriate authentication and quality checks; it does not replace them.

Establish identity at the source, then preserve it locally

At acquisition or thaw, record the source, supplier or repository identifier where available, the stock or vial reference, and the relationship to the local culture record. Add a stable local ID that remains unambiguous even when several cultures share the same descriptive name. If the material is an engineered clone or derivative, preserve that distinction explicitly.

Geraghty and colleagues’ guidelines emphasise reliable sourcing, authentication and banking of authenticated material. A record should point to the relevant evidence and its date. Recording that a test was performed is different from making the result accessible; recording an identity check is also different from establishing that the culture is free of contamination. ^[3]

Define what counts as a new local record. A new thaw, a separately maintained branch, a clone or a differentiated derivative may need its own identity even when it shares much of its ancestry with another culture. Apply the rule consistently and retain the parent relationship rather than reusing an old ID for new material.

Capture the event, not just the latest state

“Passage 12” describes a state. A dated passage event can also describe its source, its destination, the operator, and the procedure used. Both are useful, but the state alone cannot explain how two cultures at the same passage number developed different histories.

Keep planned and actual work distinct. The planned date, recorded outcome, event date and entry time answer different questions. For retrospective corrections, preserve the reason and the earlier record. Keep a protocol version or identifiable method reference with the work so that a later revision does not silently change the description of an earlier experiment.

When a culture branches, link each child to the appropriate parent and retain events specific to that branch. A treatment, material change, observation or quality concern affecting one branch should not automatically be inherited as if it occurred in all siblings. Shared ancestry and shared exposure are different relationships.

Link observations and results without losing the originals

An image becomes more interpretable when it carries the culture ID, acquisition date, vessel or sample identity, and enough acquisition context to understand what is shown. Keep the original file available under the laboratory’s storage policy. A cropped image pasted into a presentation may be useful for communication but is not the whole observational record.

The FAIR principles call for persistent identifiers, rich metadata and provenance to support reuse. At laboratory scale, that suggests a straightforward habit: give samples and files identifiers, record their relationship to the culture, and retain an understandable route from a summary result back to the source evidence. ^[4]

Do not confuse record count with experimental independence. Several images from one culture or several wells derived from one preparation may share important sources of variation. Use the lineage to help define the experimental unit for the analysis, and record that decision. A larger number of files is not automatically a larger number of biological replicates.

• Source: stock or vial, local culture ID and relevant identity evidence.
• History: thaw, passages, branch relationships, operator and dated outcomes.
• Conditions: protocol reference, material identity and documented changes.
• Evidence: observations, original images, sample IDs and relevant quality results.
• Interpretation: deviations, unresolved questions and links to the analysis or report.

Use the history to investigate, not to manufacture certainty

When two runs disagree, compare their paths before deciding which result is “wrong”. Did they begin from the same stock? Were they maintained as separate branches? Did their material or protocol histories diverge? Were the relevant observations made before or after the suspected change? The record should help you formulate questions that can be tested.

A difference that coincides with a passage or medium change is a lead, not proof of cause. Several variables may have changed together. Preserve the timeline and decide what additional evidence or comparison would distinguish the explanations. If the evidence is insufficient, record the limitation rather than polishing the history into a single confident narrative.

Culture records can also prevent avoidable confusion during reporting. State the source and relevant history actually used, rather than copying a generic cell-line description from a previous manuscript. Check that the methods, sample identifiers and available raw records describe the same experimental material.

Make connected records a routine, manageable practice

Start with the points at which information is easiest to lose: receipt or thaw, passage into new vessels, transfer of responsibility, preparation or substitution of materials, and creation of samples for downstream analysis. Agree who records each transition and when. Avoid making one person reconstruct the entire chain at the end of a project.

Keep the overview concise: identity, current state and next action. Place the detailed evidence within reach rather than crowding every daily task with the whole history. A useful interface should support both a quick operational decision and a slower scientific review.

Periodically select a sample or figure and trace it back to the culture and source stock. Then select a stock and identify its recorded descendants. Note broken links and fix the point in the process that allowed them. The measure of success is whether another researcher can follow the evidence, including its uncertainties, without depending on your memory.