Date: Sep 17, 2022

Dataset Title: Seq-Scope processed datasets for skin results (RDS) and H&E images

Dataset Creators: Hsu, Jer-En; Cho, Chun-Seok; Kim, Myungjin; Xi, Jingyue; Lee, Jun Hee

Dataset Contact: Jun Hee Lee leeju@umich.edu

Funding: J.-E.H. is supported by a Taiwan Government Scholarship. J.H.L. is supported by a Michigan Translational Research and Commercialization (MTRAC) award and NIH UG3-CA268091. C.-S.C. is supported by NIH T32-AG000114. M.K. is supported by NIH K01-AG061236.

Overview: There are three experimental outputs from Seq-Scope. (1) High definition map coordinate identifier (HDMI) sequence, tile and spatial coordinate information from 1st-Seq, (2) HDMI sequence, coupled with cDNA sequence from 2nd-Seq, and (3) Histological image obtained from Hematoxylin and Eosin (H&E) staining of the tissue slice. (1) and (2) were uploaded to GEO (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE186601). (3) is deposited here. In addition, this deposit includes the processed RDS (single R object) data files.

Methodology: A 6 mm punch biopsy from an acne papule on the back was frozen in OCT medium and stored at -80°C until sectioning. The current study utilized Seq-ScopeHISEQ, which is similar to the previously described version of Seq-ScopeMISEQ, but utilized the HISEQ2500 flow cells for Seq-Scope array generation, instead of the MISEQ flow cells. All the procedures are extensively described in the former paper, and current study was performed almost identically; however, the following things were different between the two studies: (1) To perform the 1st-Seq process, HISEQ2500 sequencing was performed with 100 pM of HDMI32-oligo library, and generated 1.1 million clusters/mm2 fully sequenced pixel density (PF clusters). (2) HISEQ2500 tiles have a rectangular shape and are arranged to form a minimally interrupted large imaging area. Therefore, images from individual tiles were assembled to present a larger image that can be aligned with the hematoxylin and eosin histology results. (3) To accommodate the larger imaging area, volume of the solutions was proportionally increased during the experiment. (4) During the secondary strand synthesis, we used an updated Randomer sequence (5’-TCAGACGTGTGCTCTTCCGATCTNNNNNNNNB-3’). Compared to the 
original Randomer (5’-TCAGACGTGTGCTCTTCCGATCTNNNNNNNNN-3’), the updated Randomer sequence does not anneal to the poly-A region, enabling more efficient transcriptome alignment. (5) To further substantiate our former results from scRNA-seq and spatial-seq, we focused on one acne sample, which presented an ideal spatial orientation that reveals the relationship between diverse inflammatory populations (including TREM2 macrophages) and hair follicle pathology. Other experimental details and analysis methods are as described in our studies (Cho et al. Cell 184, 3559–3572; Do et al. Science Immunology 10.1126/sciimmunol.abo2787).

Files contained here:

H_E.tif
>>> Raw H&E images for the Acne skin.
skin_10um_original.RDS
>>> Skin Acne dataset binned with 10um-sided square grid
skin_2um_sliding.RDS
>>> Skin Acne dataset analyzed through sliding grids with 2um intervals

Related publication(s):
Cho CS, Xi J, Si Y, Park SR, Hsu JE, Kim M, Jun G, Kang HM, Lee JH. Microscopic examination of spatial transcriptome using Seq-Scope. Cell. 2021 Jun 24;184(13):3559-3572.e22. doi: 10.1016/j.cell.2021.05.010. Epub 2021 Jun 10. PubMed PMID: 34115981; PubMed Central PMCID: PMC8238917.
	
Do TH, Ma F, Andrade PR, Teles R et al. TREM2 macrophages induced by human lipids drive inflammation in acne lesions. Sci Immunol 2022 Jul 22;7(73):eabo2787. PMID: 35867799

Related data sets in NCBI Gene Expression Omnibus (GEO) repository: Hsu J, Cho C, Kim M, Xi J, Gudjonsson JE, Lee J "Seq-Scope analysis of human acne" https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE186601

Use and Access:
This data set is made available under a Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)