Compare clones with their history intact.

Cell line development selects the biological candidate that may become the starting material for a manufacturing process. The decision draws on construct and candidate identity, clonality evidence, growth, productivity, product quality and stability. A high titre is useful evidence, but it is not a complete selection decision.

A page-based ELN is built to record what was done. It can hold a pasted results table and the name of the chosen clone, but the wells, images, passages and method versions behind each number are not records it can follow, and a missing result looks the same as one nobody needed. Development guidance expects that history to survive. Careful records of the manipulation of the cell substrate should be maintained throughout its development,¹ and the criteria used to select the cell clone for production should be described in detail.²

Seal is where the screening, expansion and stability protocols run, or it reads the results from the systems that produced them. A comparison states its basis and shows where observations are not alike, and the selection record carries the chosen lineage into cell banking and upstream development.

A protocol in Seal carries instructions, fields, calculations, linked samples and review steps. During execution the team records actual values, observations, variations and instrument references, so the lineage is carried into the next experiment rather than reconstructed at the end. The fictional passage-30 study of CL-⁠19 is incomplete: its later titre is not recorded, and the record says so rather than leaving a blank.

Neil can help prepare the configuration and analyse the results. The team tests the workflow and defines the review and approval requirements before its intended use. A later revision of the protocol does not rewrite the studies made under an earlier one.

Later samples resolve to the same candidate even when instruments or laboratories use different identifiers. If a sample cannot be traced confidently, the uncertainty stays with it rather than the result being attached to the apparent best match.

Clonality is a conclusion drawn from evidence. For recombinant products, the cell substrate is the transfected cell containing the desired sequences, cloned from a single cell progenitor, and the master cell bank is usually made directly from an initial clone or from a preliminary bank derived from one.¹ Seal keeps the source images, the imaging record and the reviewer’s conclusion as separate records, so a later reviewer can see what was seen, what was concluded and on which evidence. The fictional CL-⁠07 has its day-0 image and imaging record attached; the reviewer’s conclusion is not yet recorded.

The cultivation history is part of the identity. For continuous cell lines, culture duration is usually estimated by population doublings, subcultivations at a defined dilution or time in days,¹ so each expansion keeps its passage, dates and conditions with the candidate, and each later study shows the passage it used.

In the fictional study, three candidates were cultured at passages 10 and 30 in 15 mL fed-batch cultures under recipe UP-⁠08 v02 and harvested at day 14, recorded as CLD-⁠031 to CLD-⁠033. CL-⁠12 has the highest early titre, 4.6 g/L, but reports 3.2 g/L at passage 30, about 30% lower. CL-⁠07 changes little, from 3.8 to 3.6 g/L. The later result for CL-⁠19 has not been reported, so it stays in view without being ranked as stable. Titres use TTR-⁠02 v03; SEC monomer, 97.1 to 99.0%, uses SEC-⁠04 v02 on the early harvests only. Each value is one observation, not a replicate mean.

These observations raise questions; they do not establish genetic stability or choose a clone. A difference between passages may have several explanations, so the assessment asks the team to check the culture conditions, replicate evidence and the programme’s stability design. Missing measurements stay missing, and a repeat does not erase the original observation.

Compare productivity with growth, viability and product-quality results, and expose missing tests and differences in methods. A result from a different culture format or method version is shown as such rather than ranked beside the others.

Ask Neil which candidates deserve the next study, and it returns the comparison with its sources: each candidate’s titres by passage, the quality results that exist, the results that are missing and the methods behind them. It does not reduce different kinds of evidence to a single score unless the team has defined and justified that method.

For CL-⁠07, CL-⁠12 and CL-⁠19 the useful proposal is to keep all three in the review. Review matched culture conditions and replicate evidence for the passage comparison. Complete the product-quality panel and the later-passage result for CL-⁠19. Review clonality, lineage and stability evidence against the programme’s selection criteria. The draft does not select a clone or release a bank; everything in the middle column is preparation for the scientists’ decision.

In line with the EU’s draft GMP Annex 22 on AI, Neil is not used in GMP execution. It helps set up configuration, which your team verifies and releases under change control. Customer data is not used to train AI models, and the requester’s permissions govern what Neil can read. See how Neil fits Annex 22 and more about Neil.

Cell substrate stability has two concerns: consistent production of the intended product, and retention of production capacity in storage. For stability during cultivation, at least two time points should be examined, one with cells that have had minimal subcultivation and another at or beyond the limit of in vitro cell age for production, and that limit should be based on production cells expanded at pilot or commercial scale.¹ The limit of in vitro cell age for commercial production is assessed as part of process validation.³

A 15 mL passage study is therefore a screen, not that demonstration. It shows which candidates are worth the full study and where productivity, growth or metabolic indices change with passage; oxygen or glucose consumption and ammonia or lactate production rates are among the indices that can help where the product is hard to compare directly.¹ Seal keeps the study design, the passages tested, the results that exist and the ones still missing, so the selection assessment shows how much stability evidence it actually rests on.

The expression construct has its own evidence. Its analysis establishes that the correct coding sequence was incorporated into the host cell and is maintained during culture to the end of production. It covers copy number, insertions or deletions and the number of integration sites, and the method’s validation estimates its limits of detection for variant sequences.² Copy-number and sequence results link to the candidate and passage they describe, beside its productivity and quality results.

Keep the source versions with the assessment, so a later reviewer can see what was known when the decision was made: the candidates considered, the evidence reviewed, the criteria, the exclusions, the open questions and the rationale. The master cell bank is generally derived from the selected clone, and the cell line’s history and the production of its banks should be described in detail, including the methods and reagents used in culture, in vitro cell age and storage conditions.² Bank identity, characterisation, passage limits and restrictions then travel with every vial.

In the fictional example, CL-⁠04 was selected in an earlier programme and banked as MCB-⁠01, from which WCB-⁠02 was derived; the upstream runs USP-⁠028 and USP-⁠029 start from a WCB-⁠02 vial. Cell banks are the starting point for manufacturing a biotechnological drug substance,³ and cell bank and seed lot management keeps each generation authorised and traceable.

Clone selection and upstream process development are different jobs. Give the upstream team the selected cell source and its supporting evidence, then develop media, feeds and culture conditions in upstream development.

Start with exports or scope a permitted connection to your imaging, screening, bioreactor and analytical systems. Preserve candidate and sample identities, units, timestamps and method versions; availability depends on the source interfaces. Confluence from live-cell imaging is not a direct cell count, and a failed titre fit is not zero, so each result keeps the context that qualifies it. See all integrations.

Compare the candidates, inspect the gaps and agree the next study together. Then extend the same records into upstream development and analytical development.