sequencescape icon indicating copy to clipboard operation
sequencescape copied to clipboard

GPL-871 - quadrant metadata q0-3 (PSDTHV) [C=M, V=4*]

Open KatyTaylor opened this issue 4 years ago • 0 comments

Describe the bug Metadata relating to which parent plates were the sources for which quadrants of a 384-well plate is 0-indexed in the Beckman process, but not on the previous (Sentinel) process. It causes confusion in the lab, as if you refer to 'quadrant 2', you would mean a different one depending on whether they are 0 or 1 indexed.

Reported by Anna G:

This problem can be seen on every Beckman pick in Sequenscape – plate summaries Here is an example http://sequencescape.psd.sanger.ac.uk/plate_summaries/HT-111183 It’s at the bottom and the information is from creation of LTHR -384 RT Q

Additional Information In the Sentinel process, the metadata is created by the Sequencescape quad stamping page: See quadrant_metadata_collection method in quad_creator.rb (Sequencscape repo)

SELECT cm.key, cm.value, l.`id`, l.created_at, l.name, p.name
FROM custom_metadata cm
JOIN custom_metadatum_collections cmc ON cm.`custom_metadatum_collection_id` = cmc.`id`
JOIN labware l ON cmc.`asset_id` = l.`id`
JOIN plate_purposes p ON p.id = l.`plate_purpose_id`
WHERE p.name = 'LTHR-384 RT'
ORDER BY l.created_at DESC
LIMIT 100;

Gives outcome such as:

key value id created_at name name
Quadrant 3 DNxxxxxx1 1 23/01/2021 15:00 Plate DNxxxxxxA LTHR-384 RT
Quadrant 4 DNxxxxxx2 1 23/01/2021 15:00 Plate DNxxxxxxA LTHR-384 RT
Quadrant 1 DNxxxxxx3 1 23/01/2021 15:00 Plate DNxxxxxxA LTHR-384 RT
Quadrant 2 DNxxxxxx4 1 23/01/2021 15:00 Plate DNxxxxxxA LTHR-384 RT
Quadrant 1 DNxxxxxx5 2 23/01/2021 14:47 Plate DNxxxxxxB LTHR-384 RT
Quadrant 2 DNxxxxxx6 2 23/01/2021 14:47 Plate DNxxxxxxB LTHR-384 RT
Quadrant 3 DNxxxxxx7 2 23/01/2021 14:47 Plate DNxxxxxxB LTHR-384 RT
Quadrant 4 DNxxxxxx8 2 23/01/2021 14:47 Plate DNxxxxxxB LTHR-384 RT

In the Beckman process, the metadata is created by the Limber quad stamping page: add_stock_barcode_metadata / ‘stock_barcode_from_quadrant’ in quadrant_split_plate.rb (Limber repo)

SELECT cm.key, cm.value, l.`id`, l.created_at, l.name, p.name
FROM custom_metadata cm
JOIN custom_metadatum_collections cmc ON cm.`custom_metadatum_collection_id` = cmc.`id`
JOIN labware l ON cmc.`asset_id` = l.`id`
JOIN plate_purposes p ON p.id = l.`plate_purpose_id`
WHERE p.name = 'LTHR-384 RT-Q'
ORDER BY l.created_at DESC
LIMIT 100;

Gives outcome such as:

key value id created_at name name  
stock_barcode_q0 HT-000001 5 22/01/2021 16:54 Plate DNxxxxxxE LTHR-384 RT-Q
stock_barcode_q1 HT-000002 5 22/01/2021 16:54 Plate DNxxxxxxE LTHR-384 RT-Q
stock_barcode_q2 HT-000003 5 22/01/2021 16:54 Plate DNxxxxxxE LTHR-384 RT-Q
stock_barcode_q3 HT-000004 5 22/01/2021 16:54 Plate DNxxxxxxE LTHR-384 RT-Q
created_with_robot Pre Bravo 1 5 22/01/2021 16:54 Plate DNxxxxxxE LTHR-384 RT-Q
stock_barcode_q0 HT-000005 6 22/01/2021 14:44 Plate DNxxxxxxF LTHR-384 RT-Q
stock_barcode_q1 HT-000006 6 22/01/2021 14:44 Plate DNxxxxxxF LTHR-384 RT-Q
stock_barcode_q2 HT-000007 6 22/01/2021 14:44 Plate DNxxxxxxF LTHR-384 RT-Q
stock_barcode_q3 HT-000008 6 22/01/2021 14:44 Plate DNxxxxxxF LTHR-384 RT-Q
created_with_robot Pre Bravo 1 6 22/01/2021 14:44 Plate DNxxxxxxF LTHR-384 RT-Q

Consider: if the metadata key recorded by one or other of these pages is changed, consider whether to do a data migration to alter historical data to match.

KatyTaylor avatar Jan 26 '21 13:01 KatyTaylor