search

biocore / labadmin

Administration website for the Knight Lab
4 stars 16 forks source link

Adding generate_i5_i7_indexes #222

Closed antgonza closed 7 years ago

coveralls commented 7 years ago

Coverage Status

Changes Unknown when pulling 9f6ff59db8433f73097ae924e5bd38a11ad8d463 on antgonza:generate_i5_i7_indexes into on biocore:plate-mapper-kiwi.

coveralls commented 7 years ago

Coverage Status

Changes Unknown when pulling 9f6ff59db8433f73097ae924e5bd38a11ad8d463 on antgonza:generate_i5_i7_indexes into on biocore:plate-mapper-kiwi.

coveralls commented 7 years ago

Coverage Status

Changes Unknown when pulling 52a8688646276966b31f0665b15a661a49a3fa57 on antgonza:generate_i5_i7_indexes into on biocore:plate-mapper-kiwi.

josenavas commented 7 years ago

@tanaes do we have the layout of those indexes on their plates? and it is always the same or it changes? we need to store that information to be able to generate the echo pick list.

tanaes commented 7 years ago

@josenavas We do have the layout of those indexes on their plates. It should always be the same.

One wrinkle here is that some barcode types (e.g. nextera) the i5 and i7 are pre-mixed already in the plate, so you can't add them independently.

Others (e.g. iTru) can be added independently (this is what allows us to practically scale to tens or hundreds of thousands of unique barcodes), with i5 and i7 stored in separate plates.

Maybe we want to use a barcode_plate concept? I imagine that this could be done after barcode assignment: given the set of barcodes, choose a barcode_plate or _plates that contains those barcodes and then generate the pick list.

One more wrinkle (and maybe not worth worry about now) is that the modulo / float operation above works really well with the Echo or manually with single indexes but doesn't work well manually or in the EP-motion with pre-mixed double indexes, where use of a multi-channel pipette enforces that column-wise relationships be preserved between the barcode plate and the sample/library plate. One potential resolution for this would be to have an option to have an alternative barcode assignment that just 'stamps' the barcode plate wells onto the corresponding library plate wells.

antgonza commented 7 years ago

Note that errors are not due to the code changes ...

josenavas commented 7 years ago

@wasade are your comments addressed?

wasade commented 7 years ago

sure