Journal: bioRxiv

Article Title: Deconvolution of bulk blood eQTL effects into immune cell subpopulations

doi: 10.1101/548669

Figure Lengend Snippet: With whole blood expression and FACS data of 500FG samples, Decon-cell predicts cell proportions with selected marker genes of circulating immune cell subpopulations. Validations of Decon-cell were carried out on three independent cohorts where measurements of neutrophils/granulocytes, lymphocytes and monocytes CD14+ were available, alongside to expression profiles of whole blood. Benchmarking of Decon-cell was performed against CIBERSORT and xCell . Decon-cell was applied to an independent cohort (BIOS) to predict cell counts using whole blood RNA-seq. Decon-eQTL subsequently integrates genotype and tissue expression data together with predicted cell proportions for samples in BIOS to detect cell type eQTLs. We validated Decon-eQTL using multiple independent sources, including expression profiles of purified cell subpopulations, eQTLs and chromatin mark QTLs (cmQTLs) from purified neutrophils, monocytes CD14+ and CD4+ T cells , and single cell eQTLs results . Benchmarking of Decon-eQTL was carried out for comparison with previously reported methods which detected cell type eQTL effects using whole blood expression data, i.e. Westra method and Zhernakova, et al method ).

Article Snippet: Comparison of Decon-eQTL with other methods for detecting cell type eQTLs.

Techniques: Expressing, Marker, RNA Sequencing, Purification, Comparison

Journal: bioRxiv

Article Title: Deconvolution of bulk blood eQTL effects into immune cell subpopulations

doi: 10.1101/548669

Figure Lengend Snippet: By integrating proportions of cell subpopulations (predicted by Decon-cell), gene expression and genotype information, Decon-eQTL detect cell-type eQTLs. (A) The number of deconvoluted eQTLs in each cell type by using whole blood RNA-seq data of 3,189 samples in BIOS cohort. (B) Distribution of Spearman correlation coefficients between expression levels of deconvoluted eQTL gene and cell counts for each cell subpopulation. The deconvoluted eQTL genes show positive and statistically higher correlation (Spearman) with its relevant cell type proportions than compared to the rest (T test p value < 0.05) in an independent cohort (500FG).

Article Snippet: Comparison of Decon-eQTL with other methods for detecting cell type eQTLs.

Techniques: Gene Expression, RNA Sequencing, Expressing

Journal: bioRxiv

Article Title: Deconvolution of bulk blood eQTL effects into immune cell subpopulations

doi: 10.1101/548669

Figure Lengend Snippet: (A) Expression of eQTL genes in purified cell subpopulations from BLUEPRINT is significantly higher in its relevant cell subpopulation compared to other available cell subtypes (green for granulocyte eQTL genes showing expression for purified neutrophils; orange for monocytes; purple for CD4+ T cells; pink for B cells). (B) Differential expressed genes (Adjusted p-value ≤ 0.5) between CD4+ T cells and NK cells are significantly enriched for CT eQTLs effects on CD4+ T cells (dots in purple, Fisher exact P = 1.8×10 17 ) and NK Cells (dots in yellow, Fisher exact P = 2.3×10 18 ) respectively. (C) Deconvoluted eQTLs (FDR ≤ 0.05) show significantly larger effect sizes in the purified cell eQTLs data compared to the rest of the whole blood eQTLs for which we do not detect cell type effect, as shown for deconvoluted granulocyte eQTLs in neutrophil derived eQTLs (green); monocytes (orange); CD4+ T cells (purple).

Article Snippet: Comparison of Decon-eQTL with other methods for detecting cell type eQTLs.

Techniques: Expressing, Purification, Derivative Assay

Journal: bioRxiv

Article Title: Deconvolution of bulk blood eQTL effects into immune cell subpopulations

doi: 10.1101/548669

Figure Lengend Snippet: Deconvoluted CT QTLs show high allelic concordance compared to eQTLs from purified cell subpopulations . (A) for granulocyte eQTLs (orange), Decon-eQTL achieved an allelic concordance of 99% compared to eQTLs from purified neutrophils. Similarly, the allelic concordance were 96%and 99% for monocytes and CD4+ T cells, respectively. They are higher than those observed for whole blood eQTLs when comparing to eQTLs from purified subpopulations as shown in panel (B). Deconvoluted eQTLs show an allelic concordance of 95% for significant eQTLs obtained from single cell RNA-seq data on monocytes CD14+, B cells, CD4+ T cells, CD8+ T cells and NK cells (C).

Article Snippet: Comparison of Decon-eQTL with other methods for detecting cell type eQTLs.

Techniques: Purification, RNA Sequencing

Journal: Briefings in Bioinformatics

Article Title: Chrom-Lasso: a lasso regression-based model to detect functional interactions using Hi-C data

doi: 10.1093/bib/bbab181

Figure Lengend Snippet: Comparison between Chrom-Lasso and other software. ( A ) The total number of significant cis interactions (interactions occur within the same chromosome, FDR < 0.05) detected by different methods. ( B ) Bar plots demonstrate the number of 5C interactions detected by the tested methods in indicated replicates when comparing same number of top significant interactions (upper panels), and the proportion of interactions overlap with 5C in all interactions identified (the lower panel). ( C ) Box plots show the number of promoter–promoter interactions found in the top significant interactions (left), and the proportion of promoter–promoter interactions in all interactions identified (right). ( D ) Box plots show the number of unique eQTL SNPs involving interactions found by tested methods. ( E ) Box plots compare the number of disease-associated SNPs involving interacting loci in different replicates. The comparison was done in three GWAS categories: SNPs associated with all diseases (red), SNPs associated with autoimmune diseases (green) and SNPs associated with cancer (blue). ( F ) Correlation matrices show the correlation between different replicates (treated with MboI or DpnII) based on the number of interactions detected in each TAD.

Article Snippet: Our results showed that interactions detected by Chrom-Lasso were more likely to overlap with eQTL SNPs ( , detailed methods see Materials and methods).

Techniques: Comparison, Software