README
======

Dataset title: Cellular immune endophenotypes separating early and late-onset myasthenia gravis


Creators:
Jakob Theorell

Responsible organization: Karolinska Institutet

Description: This dataset, generated as part of a study of immunocellular biomarkers of myasthenia
gravis in blood, contains single cell gene expression, B- and T-cell receptor sequences and 10
surface markers plus hashing antibodies for batch pooling from 16 individuals with the disorder. This
dataset is used in the manuscript with the same name, by Jakob Theorell et al, 2025.

	
	Date of data collection: 2021-09-08/09
	Methods used for data collection: 10X Genomics Chromium 5´kit version 3 + NovaSeq sequencing. 
	Geographic location of collection: Stockholm, Sweden. 
	Experimental & environmental conditions of collection: Samples collected as part of a cohort study 
on patients with myasthenia gravis. 


License: Generally, this data is only available for researchers with a relevant ethical permit, due to
European personal data protection laws. The licence is CC BY public copyright license. 


Essential information for dataset reuse.
----------------------------------------
An analysis workflow to regenerate the analysis in the manuscript, including intermediate 
data files, can be found here: https://github.com/jtheorell/EO_vs_LOMG



Dataset structure
--------------------
Number of files: 192, excluding the metadata files listed below.
Folder structure: Flat
File naming conventions: See information on the fastq_file_metadata.csv file below
File formats: Only fastq for the data. For metadata, see below. 


Description of metadata files
----------------------
  - Sorting_documentation_per_donor #This folder contains pdf:s showing the result of the pre-sorting excluding dead 
    #cells and debris, before going to the 10X stage. 

  - 2021-09-09-10_Cite_seq_and_hashing_protocol.docx #This is the protocol for the cite-seq and dead cell marker staining pre-10X. 

  - 2021-10-05_VDJ_21_019-024_HTO_lib_quality.pdf #This file shows the sample quality in the 10X library phase for the cite-seq data.
  - 2021-10-07_VDJ_21_019-024_ENRT_lib_quality.pdf #This file shows the sample quality in the 10X library phase for the TCR data.
  - 2021-10-11_VDJ_21_019-024_ENRB_lib_quality.pdf #This file shows the sample quality in the 10X library phase for the BCR data.
  - 2021-10-11_VDJ_21_019-024_GE_lib_quality.pdf #This file shows the sample quality in the 10X library phase for the gene expression data.

  - fastq_file_metadata.csv #This file shows an overview of the association between the individual raw fastq files and teh different datatypes, 
    #batches and donors. 
    Variables: NGI_ID, 10X Serial ID, Batch_and_hashed_pool_number, Patients in pool, Sort_date, 10X_intermediate_steps_date, Sequencing_date, 
               Sequenser, Sample_type, Filename_lane_1_R1, Filename_lane_1_R2, Filename_lane_2_R1, Filename_lane_2_R2, Filename_lane_3_R1, 
               Filename_lane_3_R2, Filename_lane_4_R1, Filename_lane_4_R2

  - file_list.csv #This file contains a list of all files.
    Variables: file_name, size

  - hash_reference_not_used_in_our_analysis.csv #This is a standard format hashing reference file. As can been in the name, we did not use this
    #file in the downstream analyses. Instead, we only used the broader cite-seq file below, which also contains the hashing antibody codes. 
    Variables: id, name, read, pattern, sequence, feature_type

  - Patient_metadata.csv #This file gives some basic background parameters per donor, including age, sex, thymic hyperplasia association, 
    #treatment, etc. 
    Variables: Sample_ID, Database_ID, Age, Sex, Years_since_diagnosis, QMG, MG_type, Thymectomy, Years_since_thymectomy, Thymus, 
               Months_since_Ab_testing, AChR_Abs, "pyridostigmin dose, mg", Days_since_immunomodulation_start, Substance

  - readme.txt #This file. 

  - TotalseqC_TBNK_and_integrated_hash_barcode_reference.csv #This file contains the cite-seq reference, which is used both for the 
    #surface marker data, but also to separate the pooled donors, as the hashing works by staining each individual donor with a
    #beta2-microglobulin antibody coupled to a unique sequence. 
    Variables: id, name, read, pattern, sequence, feature_type, patient_batch_1, patient_batch_2

#The data in these files were subsequently processed with cellranger, in all cases with default settings.
#For the gene expression data, cellranger version 7.0.1 was used. For the other three data types, cellranger version 6.1.2 was used.