Cortical interneuron development is affected in 4H leukodystrophy 69 2 RNA sequencing Prior to RNAseq, concentration and quality were confirmed using RNA Analysis ScreenTape (Agilent Technologies, USA) on a 2200 TapeStation System (Agilent) to establish RIN scores. RNA samples of sufficient quality were processed according to manufacturer’s instructions using the TruSeq Stranded Total RNA Library Prep Kit with Ribo-Zero Human (Illumina Inc., USA), generating tagged cDNA libraries capable for high-throughput RNA sequencing. Library concentration and quality were confirmed using D1000 ScreenTape (Agilent Technologies, USA) on the 2200 TapeStation System (Agilent). Samples were run on a Illumina HiSeq2500 at SR50. Quality control was performed using FastQC, and sequencing reads were aligned to Human Genome hg38 using Bowtie2(Langmead & Salzberg, 2012) with default parameters. Aligned reads were converted into count per gene using featureCount function from Rsubread package(Liao et al., 2013) with gene annotations obtained from GENCODE v26 (https://www.gencodegenes.org/) which contains unique 63,199 genes. Reads Per Kilobase per Million (RPKM) were further computed by using rpkm function from edgeR package(Robinson et al., 2010). Genes were filtered on such with count per million (CPM) > 1 in ≥ 50% of samples per cell type which resulted in 21,542 unique genes including non-coding RNAs. Sample t-SNE map To evaluate similarity between samples, we applied the t-Distributed Stochastic Neighbor Embedding (t-SNE) to the RNA-seq expression profiles. The t-SNE non-linearly projects local similarities between samples at the cost of retaining the similarities between dissimilar samples. We first normalized RPKM; zero-mean normalization followed by log2 transformation with pseudo count 1. Then t-SNE was performed 100 times and we selected the solution with the lowest Kullback-Leibler divergence. Cell type validation with ENCODE samples Gene expression profiles of 2 iPSC samples and 2 granule cell samples were obtained from ENCODE (ENCSR722POQ for iPSCs and ENCSR313IUO for granule cells). (Consortium, 2012) For expression profiles of ENCSR722PQO, obtained read counts were converted into RPKM as described above. For ENCSR313IUO, Fragments Per Kilobase Per Milling (FPKM) was available. Genes with expression value zero in all samples were filtered out. Since the expression data is not directly comparable between data sets, we performed Spearman’s rank correlations across these 4 samples from ENCODE and samples from this study by taking intersect genes (18,719 genes in total).
RkJQdWJsaXNoZXIy MTk4NDMw