Cellular immune endophenotypes separating early and late-onset myasthenia gravis

This dataset is made up of single-cell omics files from 16 individuals with myasthenia gravis at various stages of the disease and with differing disease onset age. The samples have been acquired 8+8 on a 10X Genomics Chromium 3.0 system (https://www.10xgenomics.com/platforms/chromium) on two consecutive days and all sequenced in a NovaSeq instrument (https://www.illumina.com/systems/sequencing-platforms/novaseq.html) in the same run. For each of the two experimental days, the data has been hash-tagged and then mixed. Complicating the matter somewhat, there are three sets of files (= three 10X serial IDs) for each day, each of whom contains cells that are a hashed mixture of all eight individuals from this day. Furthermore, four different kinds of data is available for each cell: RNA sequences (GE), B-cell receptor sequences (ENRB), T-cell receptor sequences (ENRT) and RNA-tagged surface-protein binding antibodies (HTO). The last portion contains a set of 10 lineage marker antibodies as well as the hashing antibodies. Further, considering that each set of files is made up of four lanes, and each lane has one file per direction, the total number of files is 6x4x4x2=192 files. This is the abstract from the forthcoming publication: The two main subgroups of autoimmune myasthenia gravis, a neuromuscular junction disorder associated with muscle weakness, are the early and late-onset forms, defined by onset before or after 50 years of age. Both carry acetylcholine-receptor autoantibodies, but differ in sex ratios, genetics and occurrence of disease-specific thymus inflammation. By applying multimodal techniques, including deep spectral cytometric phenotyping and single cell sequencing to peripheral blood and thymocyte samples we explored the possibility to discriminate the two forms by cellular immune phenotyping. Analyzing two independent cohorts we identified distinct immunological differences driven by three main lymphocyte populations. Lower frequencies of mucosa-associated invariant T cells and naïve CD8 T cells were observed in late-onset myasthenia, suggesting enhanced immune senescence. Further, a highly differentiated, canonical natural killer cell population was reduced in early-onset myasthenia, which was negatively correlated with degree of thymic inflammation. Using only the frequency of these three populations, correct myasthenia subgroup assignment could be predicted with an accuracy of 90%. These findings identify diverging cellular immune phenotypes between early and late-onset disease, indicating distinct underlying immunopathogenic processes. Together with demographic and disease subgroup-specific features, this may improve clinical classification, in turn of relevance for channeling to interventions. The dataset consists of 192 files in fastq format, compressed with gzip, and 42 files in other formats (pdf, csv, docx, and txt). Please, see documentations files, Readme.txt and file_list.csv, for details. The total size of the dataset is approximately 450 GiB.