Blogs, Genomics

NGS Plate Map Traceability: Designing Records From Extraction Batch to Sequencing Run

If an extraction control fails after libraries have entered several pools, the wet-lab manager needs more than a list of accessions in the batch. They need to know which extracts were used, which library preparations came from them, and where those libraries were pooled and loaded. They also need to distinguish work that finished from work that was only planned.

That is a useful test of NGS plate map traceability. Can the laboratory answer those questions from the preparation records, or does someone have to open several workbooks and reconstruct the answer?

In short, NGS plate map traceability means the LIMS can follow any sample from extraction batch to sequencing run, and back again. That takes a separate record for each extract, library, pool, and loading, transfer records that separate the plan from what actually happened, and index assignments checked within each demultiplexing scope.

The difficulty often lies in the way the records are organized. An accession connects work to the case, but the same case can produce a second extract, two library preparations, or contributions to more than one pool. If the LIMS keeps only the latest well and run against the accession, those distinctions disappear.

CLIA’s specimen identification and integrity standard, 42 CFR 493.1232, requires written policies and procedures to ensure positive identification and optimum integrity from collection or receipt through testing and reporting. It does not specify a plate-tracking data model. The engineering task is to give the laboratory records that support its approved procedures and can be followed in either direction when something needs investigation.

This article stays at the plate level. The wider traceable genomics laboratory workflow, from order to report, is covered separately.

NGS plate map traceability starts with the material in the well

C03 means very little without a plate identifier and the processing context. It may contain an extract on one plate and a library on another. Even the same physical position can hold different material at different points in a workflow.

The underlying records need to distinguish the accession, received specimen, extract, library, and pool. Containers and wells describe where material is held. They should not become the material’s only identity.

That distinction matters during repeat work. A second extraction produces another extract from the source specimen; it is not a correction to the first extract’s location. A new library preparation from an existing extract is a different preparation, even though the patient and accession have not changed.

Consider this simplified DNA workflow, in which individually indexed libraries feed final sequencing pools. Other assay designs require a different sequence of steps.

Illustrative material lineage:

StepIdentifierContainer and wellComes from
SpecimenSP-318Received specimenSource
ExtractEX-318-1Extraction plate EXT-42, well C03SP-318
LibraryLB-318-APreparation plate LIB-17, well G08EX-318-1
Pool contributionPOOL-9Not shownLB-318-A and other libraries
Run loadingRUN-12Not shownPOOL-9
LibraryLB-318-BPreparation plate LIB-18, well B02EX-318-1
Pool contributionPOOL-11Not shownLB-318-B and other libraries
Run loadingRUN-14Not shownPOOL-11

Other libraries contribute to each pool; they are omitted here for clarity.

Both libraries came from one extract. They remain separately identifiable, and each has its own route into a pool and run loading. If a library contributes to a further pool, the new contribution can be recorded without overwriting the earlier one.

A relational database can represent these relationships. The important decision is not whether to buy a graph database, but whether preparation inputs, derived materials, transfers, pool contributions, and loadings are explicit enough to query.

How should a LIMS record liquid handler transfers?

A liquid-handler worklist is an instruction, not an execution record. An interrupted worklist exposes a weakness that a completed run can conceal.

Suppose the liquid handler stops halfway through a transfer. The destination plate map was generated correctly, and the right source and destination plates were scanned. None of that establishes that every destination received material.

The LIMS needs to retain the issued plan and the evidence returned from execution. Completed, failed, skipped, and unresolved transfers should remain distinguishable. The supporting record might be an instrument log, operator confirmation, or the outcome of an approved exception process; its origin should be clear.

A practical transfer record includes:

Record elementWhat the laboratory needs to recover
Source material, container, and positionWhat was intended to be transferred and from where
Destination material or preparation, container, and positionWhere the material contributes
Transfer type and applicable procedureWhether the action is an aliquot, preparation input, normalization, or pooling contribution
Quantity, unit, and evidence typeWhether the quantity was planned, commanded, reported, or measured
Worklist version, instrument, operator, and timeThe context in which the work was performed
Outcome and supporting recordWhat is known about the transfer, including uncertainty
Exception and correction historyHow a discrepancy was investigated and resolved

The quantity field deserves particular care. A programmed 2 µL transfer is not an independently measured 2 µL delivery just because the instrument completed its program. Calling both actual volume makes the record look more certain than the evidence supports.

The same problem appears during recovery. Reissuing a whole worklist must not silently mark already completed transfers as having occurred again. The recovery process needs to account for what finished, what did not, and what remains uncertain.

Manual transfers need a capture process that fits the bench. Decide which scans and confirmations belong at the point of work, and how an operator records a deviation without losing their place. A technically complete form that staff can only fill out retrospectively is a poor substitute for a workable execution record.

Bring plate tracking into your laboratory workflow software. NonStop’s sample tracking software for clinical labs and LIMS customization services cover plate mapping, instrument interfaces, and workflow configuration.

What should LIMS plate tracking record about plate layouts and controls?

A file containing the right sample identifiers can still produce the wrong physical arrangement. Plate geometry is part of the method.

Row-first filling, column-first filling, and a defined transfer between different plate formats are different mappings. A reliable implementation records the labware definition, supported addresses, orientation convention, and source-to-destination transformation. It does not reconstruct the transformation later from the order of rows in an export.

Address handling needs to be predictable too. Where A1 and A01 mean the same well, normalize them consistently and retain the source representation needed to investigate an import. Check addresses against the labware type, not a generic list of valid-looking coordinates.

Controls need their own representation. An empty well is not an extraction blank, and an extraction blank is not a no-template amplification control. Preserve the control type, process stage, linked QC result, and applicable disposition rule. Creating a patient accession simply to accommodate a control in the schema obscures what the control represents.

If a carrier identifier is reused, retain a separate usage context. Otherwise, a lookup against an old transfer can return the carrier’s current contents instead of its contents at the time of execution. Historical plate maps must be recoverable at the relevant processing point.

CAP’s NGS worksheets are presented as concepts and teaching examples, not prescriptive performance standards. The same qualification applies to this design: laboratory leadership sets control placement, QC criteria, and acceptance policy. The software must represent that policy accurately rather than supply an unapproved default.

Keep sample barcode traceability separate from index assignment

The word barcode can conceal three different records: the label scanned on a container, the index sequence associated with a library, and the well address in a plate map. A clinical NGS implementation needs to keep them separate, even if staff use shorthand at the bench.

For an indexed workflow, preserve the library’s index assignment, the index-set or adapter definition, and the reagent identity required by the procedure. With dual indexing, retain both sequences and the convention used to export them for the supported sequencing and demultiplexing configuration.

An index name alone is not an export specification. The sequence representation and orientation need to match the instrument and software combination and the laboratory’s approved configuration. Treat that mapping as something to verify, not an assumption carried over from another platform.

Index collision checks belong to the actual demultiplexing scope, including lane handling where applicable. Two libraries intended to remain separately identifiable cannot have an ambiguous assignment in that scope. Reusing an index in an unrelated historical run is not the same problem as assigning it ambiguously to libraries processed together.

There is also a limit to what correct metadata establishes. Research on index swapping and non-redundant dual indexing, including Costello et al. in BMC Genomics (2018), shows why library preparation, sequencing chemistry, and filtering remain relevant to read assignment. With unique dual indexes, swapped reads can be identified and filtered; with combinatorial dual indexes, a swapped read can still be assigned to the wrong sample. Dual indexing should not be presented as eliminating every contamination or misassignment event.

Likewise, a successful scan confirms the identifier presented to the reader. It does not independently establish the biological identity of the material inside the tube. LIMS plate tracking supports physical and analytical identity controls; it does not replace them.

Library pooling traceability: a pool needs a contributor record, not just a name

A pool identifier tells the laboratory which preparation it is handling. The contributor records explain what went into it.

For each input, preserve the source preparation, transfer quantity and unit, execution evidence, and the concentration or QC record used to calculate the planned contribution where relevant. Those records need to remain accessible after the pooling worksheet has served its immediate purpose.

Keep planned membership separate from executed or unresolved membership. If a library was omitted during pooling, its presence on the original plan must not make it an executed contributor in the LIMS.

Contribution quantities explain assembly. They do not guarantee that every library will produce a matching fraction of sequencing reads.

Repeat work then has a straightforward representation:

  • Re-extraction creates a distinct extract linked to its source specimen.
  • A genuinely new library preparation creates a distinct library linked to its preparation inputs.
  • A new pool has its own contributor records, including libraries used in an earlier pool.
  • Loading an existing pool again creates a further loading and run relationship without inventing another library preparation.

Corrections need equal discipline. Preserve the original record, the change, the reason, and the required review. Correcting a wrongly recorded well is not necessarily a physical transfer, so the application should not create a fictional transfer to explain the edit.

How do you trace an extraction control failure to affected libraries?

Return to the extraction-control example. This is an illustrative investigation, not a rule for automatic rejection.

The first question is which extracts were associated with the questioned process. From there, the records should identify derived libraries, executed pool contributions, and the corresponding loadings. A reverse query should recover the same preparation history from a library or loading record.

Keep direct descendants separate from other materials that shared a pool or processing context. Both sets may matter, but they answer different investigation questions.

A direct relationship establishes lineage. Sharing a pool may identify material that needs assessment for possible exposure. Neither observation, on its own, proves contamination or determines the appropriate disposition of a sample or run.

Qualified laboratory personnel decide what to hold, assess, repeat, or release under the applicable procedure. The software’s contribution is to make the candidate records and supporting evidence retrievable, without asking someone to infer relationships by comparing plate workbooks.

Discuss the transfers and exceptions your LIMS must support. Bring your plate formats, preparation steps, instrument worklists, and pooling rules to NonStop.

Sequencing run traceability depends on what was loaded

A pool assigned to a planned run has not necessarily reached the instrument.

Retain the loading event or attempt against the pool, run, instrument, and platform-specific loading unit. Include flow-cell or lane identifiers where the platform and workflow use them, along with the available operator confirmation or instrument evidence.

The record needs to distinguish a planned loading from one that occurred, was aborted, or remains unresolved. Reconcile the recorded loading with run metadata; the existence of a run identifier does not establish the physical event.

Generate or reconcile the sample sheet or equivalent run configuration against the approved library and index records and the applicable loading context. Keep the exported version with the run. A file regenerated later may be useful, but it must not silently replace the record of what was used.

In a NonStop custom LIMS, sequencers connect through sample-sheet generation and run-metadata ingestion. The wider question of how to connect sequencing instruments to LIMS and EHR systems is covered in its own guide.

That is the boundary of this article. The subsequent connection to analytical inputs and outputs is covered in NonStop’s wet lab and bioinformatics integration article. The physical preparation record and the analytical execution record need to connect without being treated as the same thing.

Where NonStop fits in NGS sample tracking

The intake record gives downstream work its starting identifiers. It cannot establish that every subsequent physical action occurred correctly.

CLIA’s specimen-handling requirements in 42 CFR 493.1242 include written procedures for labeling, processing, and specimen acceptability. Accessioning needs to establish the appropriate source identifiers before later preparation records can rely on them. NonStop’s CLIA software compliance checklist covers the wider 42 CFR 493 requirements for laboratory software.

SmartReq, NonStop’s requisition intake software, supports requisition extraction, schema validation, low-confidence review, and an audit log of edits, with LIMS mapping tailored during implementation. That is its role here: structured order information. It does not confirm the biological contents of a tube or verify a bench transfer.

NonStop’s custom LIMS development work covers specimen tracking, barcode configuration, workflow rules, instrument integration, and auditability. For plate-to-run traceability, the engagement should explicitly scope the material identities, transfer evidence, contributor records, and loading relationships the laboratory needs. A plate display alone does not establish that those behaviors have been implemented.

For a U.S. genetic testing leader, NonStop rebuilt sample management around a LIMS-driven architecture for traceability and modernized ordering and reporting while the platform stayed live. Turnaround time fell by 55%.

The CDC and APHL NGS Quality Initiative provides customizable quality-management resources for clinical and public health laboratories. The implementation should support the laboratory’s procedures and quality records, with acceptance criteria and quality decisions remaining under laboratory leadership.

NonStop’s forward-deployed engineering model puts implementation engineers alongside the client’s workflow through delivery. For this work, that means resolving practical questions with wet-lab and informatics staff: what the instrument really reports, what staff can capture during preparation, and how an investigator retrieves the evidence later.

Test NGS sample tracking software with the awkward cases

A clean, fully completed plate is not enough to assess the build. Ask for an interrupted worklist, an omitted pool contributor, a repeat library preparation, and a pool loaded more than once.

For each example, follow the record. The original plan should remain available. Execution should be distinguishable from intention. Each preparation should retain its source, and both directions of the trace should return the expected relationships.

Test a control without a patient accession, an unresolved transfer, and a changed plate-format mapping. Historical records should not acquire a different meaning because today’s configuration changed.

These are software acceptance scenarios, not substitutes for assay validation or laboratory approval. They belong in any LIMS vendor evaluation and help a buyer evaluate whether the sample tracking platform represents the bench process accurately before the laboratory depends on it.

Design plate-to-run traceability around your actual preparation workflow. Bring your extraction layouts, transfer records, index definitions, and pooling process to NonStop’s architecture review.

Questions about clinical NGS plate and barcode traceability

How do you track samples from extraction batch to sequencing run?

Identify each physical preparation, link it to its source, and preserve the transfers and contributions connecting extracts, libraries, pools, and loadings. Keep the execution evidence separate from the planned plate layout so incomplete work remains visible.

Does dual indexing prevent index hopping?

Not entirely. Index swapping still occurs, especially on patterned flow cell instruments. With unique dual indexes, a swapped read carries an unexpected index pair and can be filtered out. With combinatorial dual indexes, a swapped read can match another sample’s valid pair and be assigned to that sample.

What is an index collision in NGS pooling?

An index collision happens when two libraries meant to stay separate have index assignments that cannot be told apart within the same demultiplexing scope, such as a lane. That includes sequences too similar for the mismatch tolerance used during demultiplexing. The LIMS should flag the pool before it is loaded.

What should a LIMS keep in a historical plate map?

Preserve the container and well, material identity, processing context, control type, applicable plan version, and the events explaining occupancy or derivation. Historical maps should be recoverable with their supporting execution records.

Can a barcode scan prove that the correct sample was transferred?

It checks the identifier presented to the reader. It does not independently establish biological contents or every aspect of dispensing. Use scanning alongside the laboratory’s approved physical and analytical identity controls.

What should you ask a LIMS vendor about NGS plate tracking?

Ask how the software handles split preparations, multi-contributor pools, partial execution, controls, repeat work, and historical reconstruction. Specify the labware, instruments, and approved index-export conventions the implementation needs to support.