LocusLab Discuss a method change

Review changes to automated lab methods.

Compare two versions of a method. Bring each change, its source lines, the earlier test evidence and the open questions into one review.

For teams developing NGS library preparation on liquid handlers such as Hamilton STAR or Opentrons Flex.

Discuss a method change See how a review is built
Change request
CR-01
Change
2 of 4
Title
Cleanup transfer from Instrument A to Instrument B

Bead to sample ratio

Before
0.80 ratio
After
0.90 ratio

07_parameters_before.json · /parameters/0/value

9"value": "0.80",

08_parameters_after.json · /parameters/0/value

9"value": "0.90",

SYNTHETIC_T1 2026-09-29 Instrument A, ratio 0.80

T1's record declares ratio 0.80; the after parameters declare 0.90. The link records that difference and assesses nothing.

Context only Applicability: Not assessed

12_trial_T1.json · /configurationSnapshot/beadToSampleRatio

SYNTHETIC_T2 2026-10-01 Instrument A, ratio 0.80

T2's record declares Instrument A and the old ratio 0.80. Its two documents declare different bead lots, and the record names no QC table.

Context only Applicability: Unknown

14_trial_T2_record.json · /configurationSnapshot/beadToSampleRatio

Open question for SYNTHETIC Method Owner Open

Is there a trial record for Instrument B at ratio 0.90? Both supplied trials declare Instrument A at 0.80.

Reviewed by
Date
Decision

Use it when a tested method has to change.

A new sample type, a different volume, a labware swap or a move to another instrument changes the conditions a method was tested under.

Reviewing that change means lining up the two versions, finding the conditions of the earlier trials and writing down what is still unknown. Today that work is spread across scripts, worklists, run logs, spreadsheets and notes.

The example on this page starts from one such request: a bead cleanup step moves to a second instrument, one ratio changes and a label is corrected.

CR-01 The change request

Synthetic example

01_change_request.md · L12

12Request: transfer the existing cleanup step from Instrument A to Instrument B. Both have the declared model SYNTHETIC Model 1. Change bead_to_sample_ratio from 0.80 to 0.90 and correct the plate display label from Libary cleanup to Library cleanup.

How a review is built

Four steps, shown on the same synthetic change request.

  1. Add the files you already have.

    Method versions, parameter files, sample plans, trial records, QC tables and notes. Text, JSON, CSV and instrument method files are read. Other formats stay attached as originals. No original is modified.

    Sources 16

    Synthetic example
    • 01_change_request.md Record
    • 02_independent_sample_index_request.csv Record
    • 03_method_before.txt Method, before
    • 04_method_after.txt Method, after
    • 05_deck_A.json Record
    • 06_deck_B.json Record
    • 07_parameters_before.json Parameters, before
    • 08_parameters_after.json Parameters, after
    • 09_sample_index_before.csv Record
    • 10_sample_index_after.csv Record
    • 11_QC_SOP_excerpt.md Record
    • 12_trial_T1.json Record
    • 13_QC_T1.csv Measurements
    • 14_trial_T2_record.json Record
    • 15_trial_T2_operator_note.md Note
    • 16_T1_exclusion_note.md Note

  2. See each change next to its source.

    Text is compared line by line. Declared parameters are compared by key, value and unit. A reviewer can add a change from any passage. Every change keeps its origin and its exact citations.

    In the example, the fourth change was not in the request: the after-plan exchanges the indexes of two samples.

    Changes 4

    Synthetic example

    Declared transfer from Instrument A to Instrument B Manual

    Instrument A → Instrument B

    Bead to sample ratio Calculated from declarations

    0.80 ratio → 0.90 ratio

    Plate display label corrected Manual

    Libary cleanup → Library cleanup

    After-declared sample/index identities differ from the independent request Manual

    S01: IDX01; S02: IDX02 → S01: IDX02; S02: IDX01

    sample_idindex_iddeclared_volume_ul
    SYNTHETIC_S01SYNTHETIC_IDX0120
    SYNTHETIC_S02SYNTHETIC_IDX0220
    sample_idindex_iddeclared_volume_ul
    SYNTHETIC_S01SYNTHETIC_IDX0220
    SYNTHETIC_S02SYNTHETIC_IDX0120

    Contradicts Applicability: Not assessed

  3. Read earlier trials against the new conditions.

    Each trial record is linked to the change it relates to, with its declared conditions beside the target.

    Whether an earlier result applies to the new conditions is entered by a named person. Until then it reads Not assessed. Your team reviews the evidence and decides what to retest.

    Trials and target conditions

    Synthetic example
    Declared condition Trial T1, 2026-09-29 Trial T2, 2026-10-01 Target, after method
    Instrument Instrument A Instrument A Instrument B
    Bead to sample ratio 0.80 0.80 0.90
    Deck and plate position SYNTHETIC_DECK_A, position_2 SYNTHETIC_DECK_A, position_2 SYNTHETIC_DECK_B, position_3
    Deck version SYNTHETIC_DECK_A_V1 SYNTHETIC_DECK_A_V1 Not supplied Missing
    Bead lot SYNTHETIC_BEADS_L1 Record: SYNTHETIC_BEADS_L1
    Operator note: SYNTHETIC_BEADS_L2
    Contradiction    		 SYNTHETIC_BEADS_L1
    QC table 13_QC_T1.csv Not supplied Missing No trial supplied

  4. Leave with the open points written down.

    Contradictions between documents and missing information are recorded as findings, each with its citations.

    A named reviewer records the conclusion. The review can be presented and exported with its original files.

    Review summary

    Synthetic example
    Findings
    1 contradiction 6 missing
    Questions
    5 open
    Conclusion
    None recorded

    Open questions 5 open

    • Which declared version does the Instrument B deck have?
    • Is there a trial record for Instrument B at ratio 0.90? Both supplied trials declare Instrument A at 0.80.
    • Is the exchange of the S01 and S02 indexes in the after-plan intended? The request states that sample/index associations remain as in the independent table.
    • Which bead lot was used in T2? The record and the operator note disagree.
    • Does the QC owner intend a criterion for Instrument B? The supplied criteria are limited to T1 on SYNTHETIC_CFG_A_R1.

Help shape what comes next.

Next: the same review, callable by your coding agent. When an assistant edits a method, it compares the versions, pulls the earlier trials and leaves the open questions to your team.

Figure
1
Change
2 of 4
Shows
Change request CR-01 open in LocusLab
LocusLab showing the same synthetic review: the list of four changes, the bead to sample ratio change selected with its before and after values and file references, and the two earlier trials linked to it, marked Context only.

  • It reads your method files.

    Text is compared line by line, declared parameters by key, value and unit, CSV tables by column. Instrument-specific method formats are read as well.

  • Every record points at its source.

    Changes, evidence links, findings and questions cite exact lines, JSON values or CSV cells. Each record keeps its origin: calculated, manual, or proposed by a model and accepted.

  • Your team makes the call.

    Applicability, answers and the conclusion are entered by a reviewer and recorded under that name. Nothing is decided for you.

  • The review travels with its evidence.

    Present it with every origin and source still open to inspection, or export one file that holds the prepared records and the original sources.

  • Model preparation is optional.

    A model can propose items for the review, each with its citations. Code checks the citations and the structure, and a person accepts, edits or rejects each proposal. The model never decides whether evidence applies.

  • Files are read in your browser.

    There is no server to upload them to. Sources reach a model provider only if you configure one and confirm each run.

Bring one recent change.

Tell us about a change to one of your NGS methods: what changed, what you had to look up and what stayed open. We can walk through the synthetic example together.

Discuss a method change

Vivien Perrelle

Co-founder

Software engineer working on lab automation reliability and testing. Contributor to PyLabRobot, the open-source lab automation library. Hands-on background in electrochemistry, biosensors and instrumentation.