GEO Accession viewer

NCBI > GEO > Accession Display

Not logged in | Login

GEO help: Mouse over screen elements for information.

Sample GSM5655498

Query DataSets for GSM5655498

Status

Public on Feb 21, 2022

Title

Donor 2 pooled regions [CSP]

Sample type

SRA

Source name

Donor 2 pooled regions

Organism

Homo sapiens

Characteristics

individual: Donor 2
tissue: mixed sample (epithelium isolated from Duodenum, Jejunum, Ileum, Ascending Colon, Transverse Colon, and Descending Colon)
hashtag: mixed sample (TotalSeq-B0251 to TotalSeq-B0256)

Extracted molecule

protein

Extraction protocol

Full human intestinal tracts (duodenum through descending colon) were received from HonorBridge (formerly Carolina Donor Services). Intestines were transported on ice in University of Wisconsin Solution, and tissue dissection began at UNC Chapel Hill within eight hours of cross-clamping. Fat and connective tissue were trimmed, and intestines were measured and regions separated, with the proximal 20 cm deemed Duodenum, Jejunum and ileum evenly splitting the remaining small intestine, and colon split into equal thirds for ascending, transverse, then descending colon. Two 3x3 cm mucosectomies were isolated from the center of each region for dissociation. Full mucosa pieces were incubated in 10 mM NAC (Sigma-Aldrich A9165) in dPBS (Gibco 14190-144) at room temperature for 30 min to remove mucus, then tissue was moved to ice-cold Isolation Buffer consisting of 5.6 mmol/L Na2HPO4 (Sigma S7907), 8.0 mmol/L KH2PO4 (Sigma P5655), 96.2 mmol/L NaCl (Sigma S5886), 1.6 mmol/L KCl (Sigma P5405), 43.4 mmol/L Sucrose (Fisher BP 220-1), 54.9 mmol/L d-sorbitol (Fisher BP439-500), and 100 µmol/L Y27632 (Selleck Chemical, S6390) then washed several times by gently inverting the tubes. Tissues were then incubated in Isolation Buffer with 2 mmol/L EDTA (Corning 46-034-Cl) and 0.5mmol/L DTT (Fisher Scientific BP172-5), then shaken vigorously to remove crypts. Shakes were repeated several times, checking for crypts and/or villi each time. High-yield shakes were pooled for colon crypts, and small intestinal shakes were pooled to approximate 1:1 villus to crypt tissue by cell mass. Crypts and villi were dissociated to single cells using 4 mg/ml Protease VIII (Sigma P5380) in dPBS + Y-27632 on ice for ~45min with trituration via a P1000 micropipette every 10 min. Dissociation was checked on a light microscope then clumps were removed using filtration.
Single cells were washed with dPBS + Y-27632 then resuspended in Advanced DMEM/F12 (Gibco 12634-010) + 1% Bovine Serum Albumin (Fisher Scientific BP1600-1) + Y-27632 for staining with AnnexinV-APC (1:100, BioLegend 640920) and with one TotalSeq Anti-Human Hashtag Antibody per region (1:100, BioLegend B0251-B0256) to allow for tracking all six regions with a single library preparation. Cells were washed and resuspended in AdvDMEM + 1% BSA +Y-27632 for sorting on a Sony Cell Sorter SH800Z. Cells were gated using forward and backward scatter and AnnexinV to enrich for live single epithelial cells. 25k cells were collected from each separate region, then all regions were combined before sequencing. Library prep was performed with the Chromium Next GEM Single Cell 3’ GEM, Library & Gel Bead Kit v3.1 (PN-100012). Sequencing was performed on an Illumina NextSeq 500.
scRNAseq

Library strategy

OTHER

Library source

transcriptomic

Library selection

other

Instrument model

Illumina NextSeq 500

Description

csp corresponds to cell surface proteins (aka hashtags (HTOs))
clustered_annotated_adata_k10_lr0.92_v1.7.h5ad
Donor2_filtered_feature_bc_matrix.h5
mag12

Data processing

Read Alignment: single-cell fastq files were aligned to reference transcriptome GRCh38 with the 10X Cell Ranger pipeline (V4.0.0),
Downstream Processing: downstream analysis was performed with scanpy (v1.7.2).
Annotations for cell cycle phase were added following previously published methods in Moor et al 2018. Following filtering, read counts were log-transformed and normalized to the median read depth of Donor 2, which had the fewest read counts. Highly variable genes were identified with the Seurat v2 method, identifying 2777 genes that were subsequently used for principal component analysis.
Quality Filtering parameters:
Minimum genes: >500
Percent mitochondrial reads: <75%
Minimum counts: >3,000
Maximum counts: <50,000
Dimensional reduction, batch correction, Leiden clustering, and UMAP visualization: Mapped reads were filtered and counted by barcode and UMI and then transformed into an AnnData object using the Python implementation of scanpy (v1.7.2). Annotations for cell cycle phase were added following previously published methods78. The number of genes, number of UMIs, and percent mitochondrial expression for all cells in each sample were visualized and used to identify thresholds for high-quality cells to include in further analysis. Analysis of scRNAseq above quality control thresholds was carried out in Jupyter hub with python version 3.8 and Scanpy version 1.7179. Regional hashtag deconvolution followed published methods . Briefly, raw hashtag read counts were normalized using centered log ratio transformation followed by k-medoid clustering (k=6 medoids)19. Hashtag noise distributions were determined by removing the cluster with highest expression of a specific hashtag, then a negative binomial distribution was fit to the data of the remaining cells. Cells were considered positive for a hashtag if counts for the specific hashtag were above the distribution’s 99th percentile (p<0.01) threshold. Cells positive for multiple hashtags were excluded as likely doublets. For both single cell sequencing experiments no batch correction was performed as each dataset was analyzed separately. As each experiment was analyzed separately, no batch correction was performed, and the following steps were performed once for each dataset to generate an initial dataset. The number of genes, number of UMIs, and percent mitochondrial expression for all cells in each sample was visualized (Fig. S1F). The pp.normalize_total, pp.log1p, pp.regress_out, pp.highly_variable_genes, and pp.scale Scanpy functions were used to normalize, log-transform, regress out variation, find highly variable genes (HVGs), and scale and center the dataset at zero. HVGs were used to calculate PCs by using the tl.pca function in Scanpy (Fig. S1G). For Leiden clustering the Scanpy functions pp.neighbors, tl.leiden, and tl.paga were used to compute nearest neighbors, calculate Leiden clusters, and generate initial PAGA layouts for UMAP projections80,81. Clustering parameters used were as follows: Leiden resolution = 0.9, num_neighbors = 10, num_pcs = 40. Scoring gene lists was accomplished by using the scanpy.tl.score_genes function with default parameters.
Genome_build: GRCh38
Supplementary_files_format_and_content: h5 - raw feature counts for each dataset
Supplementary_files_format_and_content: h5ad - final processed and integrated dataset for figure generation and analysis

Submission date

Oct 25, 2021

Last update date

Oct 12, 2022

Contact name

Jarrett Bliton

E-mail(s)

jarrett@altisbiosystems.com

Organization name

Altis Biosystems

Street address

6 Davis Drive