Update generation of Preparation Sheets contain correct columns/values

[AmandaBirmingham]

Raised in amplicon walkthrough 07/30/2018
Note: below, "(empty)" means this column is intentionally left empty. "(empty)" is NOT part of the column name.

sample_name
BARCODE
PRIMER
Primer_Plate
Well_ID
Plating
ExtractionKit_lot
Extraction_robot
TM1000_8_tool
Primer_date
MasterMix_lot
Water_Lot
Processing_robot
TM300_8_tool
TM50_8_tool
Sample_Plate
Project_name
Orig_name
Well_description
EXPERIMENTAL_DESIGN_DESCRIPTION
LIBRARY_CONSTRUCTION_PROTOCOL
LINKER
PLATFORM
RUN_CENTER
RUN_DATE (empty)
RUN_PREFIX (empty)
pcr_primers
sequencing_meth
target_gene
target_subfragment
center_name
center_project_name (empty)
INSTRUMENT_MODEL (empty)
runid (empty)

[qiyunzhu]

Hi @AmandaBirmingham , is this something I can tackle? Could you let me know how I can start with? Thank you!

[AmandaBirmingham]

Hi @qiyunzhu : This issue will require modifying the generate_prep_information method in labman/db/process.py , specifically the part for Amplicon prep info generation. You can see an example of what the current output looks like in the TARGET_EXAMPLE constant in labman/db/tests/test_process.py . However, the wet lab indicates they are expecting it to have a different set of column names/order, etc than it currently creates. I will email you the correct example directly.

[qiyunzhu]

Hi @AmandaBirmingham I explored the header names and still wasn't sure how to start with. First, I wonder if there is a mapping between old and new names? I guess MasterMix_lot is identical to master_mix, but I don't know if kingfisher_robot and Processing_robot are the same thing. Second, the new list of header names is longer and new columns need to retrieved by modifying the SQL command. It will be great to know where they are stored in the database. Some seem to be hardcoded I guess (e.g., target_gene == 16S rRNA). My limited understanding of SQL language and labman database structure may make this task bit challenging. So I probably will need lots of help from you...

[qiyunzhu]

This is the best I can guess. Many columns are missing from the original report.

new	old
sample_name	sample_name
BARCODE	primer_set_composition?
PRIMER	marker_gene_primer_set.linker_sequence + marker_gene_primer_set.primer_sequence (per Gail Ackermann: "Here are the linkerprimer values - note these go into the column named 'PRIMER" for the prep info file")
Primer_Plate	Number at the end of primer plate name, e.g. 1 for a primer working plate based on "EMP 16S V4 primer plate 1"
Well_ID	well
Plating
ExtractionKit_lot	extraction_kit
Extraction_robot	equipment.external_id for epmotion_robot and kingfisher_robot, joined with underscore(s)
TM1000_8_tool	epmotion_tool?
Primer_date
MasterMix_lot	master_mix
Water_Lot	water_lot
Processing_robot	gdata_robot
TM300_8_tool	epmotion_tm300_8_tool
TM50_8_tool	epmotion_tm50_8_tool
Sample_Plate	plate
Project_name
Orig_name
Well_description
EXPERIMENT_DESIGN_DESCRIPTION
LIBRARY_CONSTRUCTION_PROTOCOL
LINKER	marker_gene_primer_set.linker_sequence
PLATFORM	platform
RUN_CENTER
RUN_DATE	leave empty
RUN_PREFIX	run_prefix? or leave empty?
pcr_primers
sequencing_meth
target_gene
target_subfragment
center_name
center_project_name	center_project_name? or leave empty?
INSTRUMENT_MODEL	leave empty
runid	leave empty

[AmandaBirmingham]

Thanks to @qiyunzhu 's investigation and subsequent conversations with Gail and Greg, it is clear that the current preparation sheet generation code is producing outputs that are quite different in format and content from what the current known-good processes generate. It will be necessary to non-trivially alter the code in process.py to address this.