Journal: Bioinformatics (Oxford, England)

Article Title: SimBu: bias-aware simulation of bulk RNA-seq data with variable cell-type composition.

doi: 10.1093/bioinformatics/btac499

Figure Lengend Snippet: Fig. 2. Comparison of gene-wise mean and variance of count data in different bulk and pseudo-bulk samples. The first column contains data of two true bulk datasets, Petitprez (Picelli et al., 2014) and Chen (Chen et al., 2017). Columns 2 and 3 contain simulated pseudo-bulk data based on mouse spleen scRNA-seq data (Tabula Muris), generated with the 10X Chromium (10X) and Smart-seq2 (SS2) assay, respectively. Pseudo-bulk samples were generated using only those cell types present in both the true bulk sample and the single-cell dataset, mimicking the sequencing depth of the true bulk samples

Article Snippet: In this work, we first demonstrate that SimBu can use Smartseq2 or 10X Genomics scRNA-seq data to generate pseudo-bulk RNA-seq data that faithfully reflects the biological and statistical features of true bulk RNA-seq data.

Techniques: Comparison, Generated, Sequencing

Journal: Genome Biology

Article Title: A new IDH-independent hypermethylation phenotype is associated with astrocyte-like cell state in glioblastoma

doi: 10.1186/s13059-025-03670-y

Figure Lengend Snippet: CGIs are associated with distinct chromatin signatures. A Illustration of the NMF-based decomposition applied to all CGIs (rows of V) characterized by their respective features (columns of V). Using k (signatures), the product of W and H is approximately equal to V and can be used to reconstruct it. B Heatmap representation of the H matrix exposure values by epigenetic modification (as columns) for each NMF-derived signature (as rows). The color indicates the exposure value of a given feature to each signature. Both glioblastoma (in dark gray) and NP (in light gray) samples are shown. C Heatmap representation of the W matrix values according to the NMF-derived signature across all CGIs (left) and comparison of the W matrix values with the DNA methylation levels in the glioblastoma subtypes and in the NPs. Higher values (yellow) on the W matrix heatmap indicate that one CGI is more likely to be associated with one signature. RTK2- and IDH-CIMP are also represented in black on the middle annotation bar. D Fractions and numbers of RTK2-CIMP (blue tones) and IDH-CIMP (yellow tones) CGIs associated with each signature or assigned to more than one signature. The global assignment fractions of all CGIs are shown on the left (in gray tones)

Article Snippet: In glioblastoma, scRNA-seq data (Smartseq2) from IDH-wild-type tumors were obtained from the Broad Institute Single-Cell Portal [ ].

Techniques: Modification, Derivative Assay, Comparison, DNA Methylation Assay

Journal: Genome Biology

Article Title: A new IDH-independent hypermethylation phenotype is associated with astrocyte-like cell state in glioblastoma

doi: 10.1186/s13059-025-03670-y

Figure Lengend Snippet: Cell state heterogeneity in glioblastoma is linked to CIMP. A Two-dimensional representation of the four cellular states and RTK2-CIMP expression scores of each cell. The colors indicate the expression of genes affected by RTK2-CIMP in the cells. Quadrant coordinates are based on the expression of gene modules associated with each IDH-wild-type glioblastoma cell state as published . B Cell proportions in glioblastoma tumors as estimated by ordinary least squares deconvolution and based on the known gene modules associated with each cellular state. Each cell state is represented in color, and phenotypes are indicated on the x -axis. C (top) Distribution of the mean DNA methylation values of CGIs (points) found in the promoter or within the genes assigned to each cellular state module in the RTK2 subtype tumors. CGIs with a high DNA methylation level (beta above .75) in the NPC module are labeled. For readability, a label is only shown on the CGI with the maximum beta value by gene if multiple CGIs are annotated to the same gene. (bottom) Mean gene expression (log-transformed TPM+1) of genes from the modules in the RTK2 subtype tumors. Annotated genes found on the DNA methylation plot are labeled. D Proposed model of the effect of CIMP on RTK2 tumors, explaining the high proportion of cells in the AC-like state (wider arrow). Transitions into OPC- and NPC-like states might be halted by CIMP. The transition to a MES-like state appears to be possible, as its suppression is likely not linked to DNA methylation alterations

Article Snippet: In glioblastoma, scRNA-seq data (Smartseq2) from IDH-wild-type tumors were obtained from the Broad Institute Single-Cell Portal [ ].

Techniques: Expressing, DNA Methylation Assay, Labeling, Gene Expression, Transformation Assay

Journal: Nature Immunology

Article Title: A genetic disorder reveals a hematopoietic stem cell regulatory network co-opted in leukemia

doi: 10.1038/s41590-022-01370-4

Figure Lengend Snippet: (a) Schematic of the MECOM locus annotated with the location of sgRNAs (sg1-sg9) tested for efficiency of MECOM editing. The binding site of sg8 (underlined) which is used in subsequent studies is shown, and clinical mutations annotated with amino acid number that have been described in MECOM haploinsufficient bone marrow failure (red) are indicated. (b) Predicted partial protein structure of the MECOM zinc finger domain with mutated residues shown as spheres. These mutations are expected to disrupt the structure of the zinc finger, either through abrogation of Zn coordination (H751, C766) or tethering between the ZnF (R750, R778). (c) Percent modified alleles (left y-axis) and percent LT-HSCs of total live cells (right y-axis) after CRISPR editing of primary human CD34 + HSPCs. Editing efficiency was detected at 72 hours after RNP delivery of Cas9 and sgRNA by nucleofection and percent of live cells that remained in the LT-HSC gate was evaluated on day 6. LT-HSCs are defined by the following immunophenotype: CD34 + CD45RA − CD90 + CD133 + EPCR + ITGA3 + . sg2, sg5, sg7, sg8 are sgRNAs targeting MECOM as described in Extended Data Fig. 1a. n = 3 biologically independent samples. Mean is plotted and error bars show s.e.m. (d) Comparison of Sanger sequencing followed by ICE analysis and Next Generation Sequencing (NGS) for the detection of CRISPR edits. AAVS1 (blue) and MECOM (red) edited samples were analyzed by ICE and NGS in parallel. (e) MECOM editing in human CD34 + HSPCs after RNP delivery by nucleofection. Editing frequency was detected at 48 hours by Sanger sequencing of genomic DNA. Transcription of edited MECOM alleles was determined by cDNA synthesis followed by Sanger sequencing of RNA from bulk HSPCs at 48 hours. n = 3 biologically independent samples. Mean is plotted and error bars show s.e.m. (f) MECOM expression following CRISPR editing. MECOM expression (normalized to GAPDH ) in bulk HSPCs was detected by qRT-PCR ( n = 3 AAVS1, n = 9 MECOM; three biologically independent experiments) and was normalized to expression in the AAVS1-edited sample on the same day. Mean is plotted and error bars show s.e.m. Two-sided Student t -test used. * P = 1.7e-3, ** P = 2.5e-4. (g) MECOM expression in LT-HSCs. MECOM expression (normalized to GAPDH ) was detected by qRT-PCR ( n = 3 per group; three biologically independent experiments) in bulk CD34 + HSPCs and in LT-HSCs sorted on day 3 after CRISPR editing. Mean is plotted and error bars show s.e.m. Two-sided Student t -test used. * P = 5.1e-3, ** P = 8.3e-4. (h) Expansion of LT-HSCs in culture. HSPCs were cultured in the presence ( n = 2) or absence ( n = 2) of the HSC self-renewal agonist UM171. Percent of LT-HSCs was determined by FACS as in Fig. and was used to calculate the total LT-HSC number. Cells were supplemented with fresh media every 2 days. (i) Expansion time course of bulk CD34 + HSPCs following CRISPR editing. HSPCs were thawed into HSC media containing 35 nM UM171 and underwent CRISPR editing 24 hours later. Cells were counted daily by trypan blue exclusion starting on day 2 after CRISPR editing and media was added to maintain equal confluency. n = 3 per group. Mean is plotted and error bars show s.e.m. Error bars that are shorter than the size of the symbols have been omitted for clarity. Two-sided Student t -test used. * P = 5e-3. (j) Stacked bar graph of cell cycle status of bulk HSPCs and HSC (HSC: CD34 + CD45RA − CD90 + CD133 + ) as determined by Edu incorporation and 7-AAD staining. On day 5 after CRISPR editing, cells were incubated with Edu for 2 hours, then fixed and permeabilized prior to 7-AAD and cell surface staining. AAVS1 -edited (A) and MECOM -edited (M) samples, were compared by the proportion of cells in G0/G1 (Edu − /2n DNA content), S (Edu + ), or M (Edu − />2n DNA content) in bulk CD34 + cells or CD34 + CD45RA − CD90 + HSCs. n = 3 per group. Mean is plotted and error bars show s.e.m. Two-sided Student t -test used. * P = 8.1e-3. (k) Stacked bar graph of cell cycle status of LT-HSCs as determined by transcriptional signatures of single-cell LT-HSCs. UCB CD34 + underwent CRISPR perturbation of MECOM or AAVS1 and were maintained in HSC media. On day 4 after editing, LT-HSCs were sorted and 10x scRNA sequencing was performed. There was no difference in cell cycle state in LT-HSCs following AAVS1 or MECOM editing. (l) Analysis of cell expansion following CRISPR editing. AAVS1 or MECOM edited HSPCs were labeled with CFSE and successive generations of cell divisions were determined by CFSE signal intensity on day 5 which was used to calculate the replication index, showing the total number of divided cells/cells that underwent at least one division. Mean of three independent experiments is plotted and error bars show s.e.m. Two-sided Student t -test used. * P = 5e-2. (m) Total number of LT-HSCs following MECOM editing. Primary human CD34 + HSPCs underwent CRISPR editing on day 1 after thawing and were cultured in HSC media containing UM171 which was changed every 2 days. On day 6 after editing, the percentage of immunophenotypic LT-HSCs determined by flow cytometry, and the total cell number determined by trypan blue exclusion were used to calculate the total number of LT-HSCs in culture. Mean of three independent experiments is plotted and error bars show s.e.m. Two-sided Student t -test used. * P = 4.7e-3. (n) Stacked bar plots of colony-forming assay comparing MECOM edited HSPCs derived from peripherally mobilized CD34 + cells from healthy adult donors. ( n = 6) to AAVS1 -edited controls ( n = 3). CFU-GEMM, colony-forming unit (CFU) granulocyte erythroid macrophage megakaryocyte; CFU-GM, CFU granulocyte macrophage; CFU-M, CFU macrophage; CFU-G, CFU granulocyte. Mean colony number is plotted and error bars show s.e.m. Two-sided Student t -test used. * P = 3.9e-2, ** P = 2.5e-4, *** P = 1.7e-5, ns=not significant. (o-p) NGS of MECOM in human HSPCs following CRISPR editing, prior to xenotransplantation (o) , and after harvest from bone marrow at 16 weeks of one representative mouse (p) . Sequences present at frequencies >0.5% are displayed. (q) Analysis of bone marrow of mice at week 16 following transplantation of MECOM -edited ( n = 5) and AAVS1 -edited ( n = 3) adult HSPCs. Mean is indicated by black line and each data point represents one mouse. Two-sided Student t -test used. * P = 3.8e-2. (r) Analysis of the MECOM locus of human cells harvested from mice following primary or secondary xenotransplantation. Half of the primary recipient mice (4/8) had human chimerism >0.25% (circles) and the other half had chimerism <0.25% (triangles) but had human MECOM sequences that were detectable by PCR. All of the secondary recipients had human chimerism <0.25% but had human MECOM sequences that were detectable by PCR.

Article Snippet: All sequencing data are deposited in National Center for Biotechnology Information Gene Expression Omnibus under Super Series GSE175521 , including GSE175515 for MUTZ-3 and primary human CD34 + LT-HSPC bulk RNA-seq; GSE175516 for LT-HSPC 10x Genomics single-cell RNA-seq data; GSE175518 for primary human CD34 + LT-HSPC Amplicon-seq data; GSE175520 for primary human CD34 + LT-HSPC SmartSeq2 data; GSE214399 for CTCF in MUTZ-3 ChIP-seq data; and GSE216225 for F36P, HNT34 and primary human CD34 + HSPC bulk RNA-seq data and HSPC 10x Genomics scRNA-seq data.

Techniques: Binding Assay, Modification, CRISPR, Comparison, Sequencing, Next-Generation Sequencing, cDNA Synthesis, Expressing, Quantitative RT-PCR, Cell Culture, Staining, Incubation, Labeling, Flow Cytometry, Derivative Assay, Transplantation Assay

Journal: Nature Immunology

Article Title: A genetic disorder reveals a hematopoietic stem cell regulatory network co-opted in leukemia

doi: 10.1038/s41590-022-01370-4

Figure Lengend Snippet: a , Uniform Manifold Approximation and Projection (UMAP) plot and cell type clustering of human HSCs after CRISPR editing. UCB CD34 + cells underwent CRISPR editing and were sorted 3 d later for CD34 + CD45RA - CD90 + HSCs followed by scRNA-seq. Cells were clustered by transcriptional signatures using Celltypist . CMP, common myeloid progenitor; MEMP, megakaryocyte-erythroid-mast cell progenitor; cMEMP, cycling MEMP; MEP, megakaryocyte-erythroid progenitor; cMPP, cycling multipotent progenitor; Ery, early erythroid progenitor; MK, early megakaryocyte progenitor; Eo/baso, eosinophil/basophil progenitor; Macro, macrophage progenitor; Mast, mast cell progenitor. b , UMAP plot of CD34 + CD45RA - CD90 + HSCs stratified by CRISPR edits, showing the depletion of HSCs following MECOM perturbation. AAVS1 -edited sample highlighted in red (left). MECOM -edited sample highlighted in red (right). Each sample is the combination of two biological replicates. c , Bar graph showing the number of cells in the HSC cluster in AAVS1 - and MECOM -edited samples. Mean is plotted and each of two biological replicates is shown. Total number of cells profiled in each group was 19,375 (AAVS1) and 19,821 (MECOM). d , UMAP plot of CD34 + CD45RA − CD90 + HSCs following CRISPR editing ( AAVS1 -edited (left), MECOM -edited (right)), colored according to expression of HSC signature ( CD34 , HLF and CRHBP ). e , Bar graph showing the number of cells expressing the three-gene HSC signature. An HSC signature score >0.5 indicates high expression. Mean is plotted and each of two biological replicates is shown. Total number of cells profiled in each group was 19,375 (AAVS1) and 19,821 (MECOM). f – h , UMAP plots of CD34 + CD45RA − CD90 + CD133 + EPCR + ITGA3 + LT-HSCs following CRISPR editing, indicating enrichment of the HSC signature as determined by scRNA-seq using the 10x Genomics platform ( f ), overlap of AAVS1 -edited and the MECOM -edited cells, sequenced using the 10x Genomics platform ( g ) and distribution of cells with monoallelic MECOM edits determined by G&T sequencing by SmartSeq2, compared to AAVS1 -edited cells and LT-HSCs from f ( h ).

Article Snippet: All sequencing data are deposited in National Center for Biotechnology Information Gene Expression Omnibus under Super Series GSE175521 , including GSE175515 for MUTZ-3 and primary human CD34 + LT-HSPC bulk RNA-seq; GSE175516 for LT-HSPC 10x Genomics single-cell RNA-seq data; GSE175518 for primary human CD34 + LT-HSPC Amplicon-seq data; GSE175520 for primary human CD34 + LT-HSPC SmartSeq2 data; GSE214399 for CTCF in MUTZ-3 ChIP-seq data; and GSE216225 for F36P, HNT34 and primary human CD34 + HSPC bulk RNA-seq data and HSPC 10x Genomics scRNA-seq data.

Techniques: CRISPR, Expressing, Sequencing