Computational protocol: Progressive alterations in multipotent hematopoietic progenitors underlie lymphoid cell loss in aging

Similar protocols

Protocol publication

[…] LMPPs isolated from pooled 4-mo or 14-mo mice were resuspended at a concentration of 200 cells/µl and loaded onto the C1 Single-Cell Auto Prep Integrated Fluidic Circuit for capturing 5–10 µm cells (Fluidigm), with loading buffer including LIVE/DEAD Viability/Cytotoxicity staining (Thermo Fisher Scientific). Images of captured cells were collected with a microscope (AxioObserver.Z1; ZEISS) in brightfield, GFP, and CY3 channels using ZenPro software. Single-cell RNA extraction and mRNA amplification were performed following manufacturer’s recommendations including exogenous spike-in controls. cDNA was quantified using the Quant-iT PicoGreen dsDNA Assay kit (Thermo Fisher Scientific) and a high-sensitivity DNA chip (Agilent Technologies) and was diluted to a final concentration of 0.15–0.3 ng/µl. Diluted cDNA reaction products were used to generate scRNA-seq libraries using the Nextera XT DNA Sample Preparation kit (Illumina). After PCR amplification, all samples were pooled, purified, and quantified using a high-sensitivity DNA chip. Libraries were paired-end sequenced using a HiSeq 2500 sequencing system (Illumina). We sequenced 96 libraries per rapid-run flow cell, generating ∼15 × 106 reads per library. An index for RNA-seq by expectation maximization (RSEM) was generated on the basis of the Ensembl mm10 transcriptome downloaded from the University of California, Santa Cruz Genome Browser (23,637 total genes). Read data were aligned directly to this index using RSEM/bowtie. Quantification of gene expression levels in tags per million (TPM) for all genes in all samples was performed using RSEM (v1.2.8). Before all subsequent analyses, we filtered the data as previously described (). First, we filtered out cells with <2,500 genes with log2(TPM + 1) > 2. Second, we excluded genes that were log2(TPM + 1) < 4 in aggregated data. Third, we centered the data by subtracting from each gene its mean expression (log2[TPM + 1]) across all cells. After filtering, our dataset included 94 single cell transcriptomes, 54 representing 4-mo LMPPs and 40 representing 14-mo LMPPs, and 1,467 genes. These libraries had <20% of counts mapping to mitochondrial genes (4.64 ± 3.13%; mean ± SD). Principal component analysis was performed using the R stats prcomp() function with variables scaled for unit variance and a 0.05 tolerance. GSEA was performed with the javaGSEA application (version 2.0.14) with the default settings (). Enrichment was considered significant if FDR < 25% and P < 0.05. Cell cycle genes were defined as those cycling (G1/S, S, and G2/M) in synchronized HeLa cells (; ). We refined this list by only including genes present in our filtered dataset. We plotted the mean of the G1/S transition signature versus the mean of the S + G2/M signatures for each cell. Lineage-specific gene sets (CLP, PreGM, MkP, and preCFU-E) were previously published (). Comparison with bulk RNA-seq libraries was performed by calculating the geometric mean of all 4-mo or 14-mo LMPP scRNA-seq data (log2[TPM + 1]) and comparing with the bulk RNA-seq log2(TPM + 1). Pearson correlation coefficients were calculated based on linear fit. [...] Four independent biological replicates of bulk LMPPs (20,000–32,000 cells; Lin−Sca-1+c-Kit+Flk2+CD150−CD34+), GMPs (30,000–61,000 cells; Lin−Sca-1−c-Kit+CD150−CD34+CD16/32+), or CLPs (8,500–28,142 cells; Lin−Sca-1loc-KitloFlk2+CD127+) isolated from 4-mo C57BL/6J female mice were sorted directly into 350 µl RLT buffer (QIAGEN) and flash frozen. Total RNA was isolated (QIAGEN), including DNase treatment. RNA was processed using an Ovation RNA-Seq kit (V2; NuGen). After shearing, a TruSeq DNA sample prep kit (v2; Illumina) was used to prepare libraries. Libraries were sequenced on the HiSeq 2000 platform (Illumina) at a sequencing depth of >35 million reads per sample. Transcript abundances were estimated for each RNA-seq sample using RSEM. Read counts estimated for each gene by RSEM were given as input to the R package edgeR for differential expression analysis (). Genes were considered differentially expressed among LMPPs, GMPs, and CLPs based on log fold change >2 and FDR <0.05 criteria. […]

Pipeline specifications

Software tools RSEM, edgeR
Application RNA-seq analysis