EGAS00001004344 - LungCellAtlas2020

[ofanobilbao]

Project short name:

HumanLungCellAtlas2020

Primary Wrangler:

Ami --> Ida

Secondary Wrangler:

Arsenios

Associated files

• Google Drive: folder spreadsheet

Ami's note: Response from authors about bulk data: "The immune cells for bulk mRNA sequencing were from an entirely different source (the blood was purchased from https://allcells.com/).". I have decided to not include this bulk data, as it is not from the same sample types as the single cells.

I have also decided to not include the mouse data from Tabula Muris Senis (authors also let us know that almost all the mouse data is from that datasets). I have created a separate project for Tabula Muirs Senis in ingest as the main publication is separate.

Link to Ingest:

• https://contribute.data.humancellatlas.org/projects/detail?uuid=6936da41-3692-46bb-bca1-cd0f507991e9

Published study links

• Paper: A molecular cell atlas of the human lung from single-cell RNA sequencing

• Accessioned data: EGAS00001004344

Key Events

• [x] Convert published metadata to HCA spreadsheet
• [x] Manually curate dataset to meet HCA metadata standard
• [x] Collect any matrix and cell-type annotation files
• [x] Upload sheet to validate metadata
• [ ] Check linking using ingest graph validator
• [ ] Transfer raw files to ingest to validate data files
• [ ] Ask the Secondary Wrangler for an end-to-end review of the project. Ask the Expertise Wrangler to review specific tabs if needed
• [ ] Submit dataset to Production
• [ ] Complete the Export SOP
• [ ] Convert project data to SCEA format following the SCEA conversion SOP if appropriate

[ami-day]

Request new cell type ontology "immune cell": https://github.com/HumanCellAtlas/ontology/issues/120

[ami-day]

Challenging dataset: way too many cell barcodes to link to either new cell suspensions (like an experiment accession) or an analysis file in a single cell in excel or google sheets. We need to find a programmatic way to do the linking in ingest.

[ami-day]

Note about cell type ontologies: bvarner-ebi and addiehl suggest rather than 'immune cell', I could use 'leukocyte' and that "immune cell" should be added as a related synonym to the ontology term.

[ami-day]

Found a workaround to link all cell suspensions

[ami-day]

Emailed the authors about mouse single cell seq data and bulk data.

[ami-day]

Info from Kyle: "Ah I think I understand where the confusion is coming from. The accession in the paper only includes FASTQs for 608 SmartSeq2 cells specific to the study. The remainder are from the Tabula Muris Senis are available on that data portal. Of the 608, there are 1 set of FASTQs for the “Tbx4-Cre > ZsGreen1” cells (tagged with KJT) and 2 (across 2 sequencing lanes) sets for the “Axin2-CreER > mTmG” cells (tagged with ANN). 224 KJT + 2 * (384 ANN) = 992 sets of FASTQs. The cell IDs have the form “P-16-CGCTCAGT-TTATGCGA_ANNS256” (after concatenating the lanes). This represents [The 384 well it was sorted into][The i5 and i7 indices][the sorter, Ahmad versus myself][the bcl2fastq index during de-multiplexing]."

[ami-day]

Linking data files to samples in spreadsheet: on-going

[ami-day]

Requested NCBI data delivery (fastq)

[ami-day]

Also waiting on test data: cell suspension as input to cell_suspension

[ami-day]

504bec8b-733e-41d1-a5c3-5b19289036cd

[ami-day]

Dataset exceeds maximum cell suspension count - error in ingest. Creating a ticket ebi-ait/dcp-ingest-central#869 for this.

[ami-day]

@ESapenaVentura will be working on the technical side of this with @amnonkhen

[idazucchi]

What I know so far of this dataset	Organism	Techinque	Specimens	Cell suspensions	Selection method	Sequence file	Analysis file
Human	10x	7	17	MACS		77
Human	SS2	8	9409	MACS + FACS		1
Mouse	10x	16	63			1
Mouse	SS2	2	608 - only 525 passed QC but fastqs are available for all cells
Human		9	9426
Mouse		18	671 at most

Mouse 10x data actually comes from Tabula Muris Senis

[idazucchi]

@idazucchi and @ESapenaVentura will discuss this today

[idazucchi]

I updated the metadata for human 10x and uploaded the analysis files to the hca-util area Working on human SS2 data --> too many cell suspensions to link to on analysis file, it hits the carachter limit for excel --> workaround is pooling the cell suspensions based on the plate (so waiting on #927) For the protocol I plan on applying the live cell selection but I welcome other suggestions

[idazucchi]

Issues

• Mouse smartseq2: the cell ids from the metadata don't match those in the fastqs
• Human smartseq2: the enrichment information is not clear, for several cells I'm not sure which markers were used

I am sending an email to sort out this information, hopefully I will get a reply soon

[anu-shiva]

@idazucchi Contacting the collaborators again

[ofanobilbao]

This is not on the high priority list, so while we wait it could be down prioritised

[ofanobilbao]

@idazucchi to chase one last time and close if no response

[idazucchi]

I emailed the authors again, if I don't get a reply in one month I'll close this ticket

[idazucchi]

we got a reply! I'll work on this dataset from next week, I'm trying to close the ones I have open at the moment

[idazucchi]

Open questions for review

• Mouse ss2: there is no exact ontology term for the mouse strain - just text? request new ontology term? we don't know where the mice came from so accurate annotation is impossible
• Mouse 10x: mouse_droplet_TMS_UMIs.csv is actually just a control dataset, it was not generated for this study
• Human ss2: not sure of how to represent gating NA and CD45+ Epcam- Unclear what type of cell was selected. For NA I think no FACS was done

[idazucchi]

mouse ss2 the cell ids from the metadata csv have been truncated: to match them up to the sequence file you need to discard the decimal digit for the plate well row

[arschat]

Start sec reviewing it.

[arschat]

Very well done Ida, on such an extensive dataset!

Donor

• We have some information for donor_organism.medical_history.smoking_history but not exact number of cigarettes per day. Maybe we could just provide that patient 1 has remote history of smoking while patients 2,3 are non-smokers. The schema allows it but I am not sure about the consistency of our data (given that all other datasets have the guideline format).

Cell suspension

• You could add a small explanation in the description of the pooled cell plates (row 9432 to 9485)

Analysis files

• Genome version for barcodes and possibly feature tables, could be Not applicable.

All file and biomaterial mappings are verified. Awesome work!

[idazucchi]

exported and filled the import form!

[idazucchi]

verified in the browser