Additional file 2 of Copy number normalization distinguishes differential signals driven by copy number differences in ATAC-seq and ChIP-seq

Additional file 2: Supplementary Fig. 1. Copy number variation drives the differential signals in ATAC-seq data upon using edgeR. (a) Trends in differential chromatin accessibility biases caused by copy number differences, shown without (left) and with (right) copy number (CN) normalization. Yellow lines represent the generalized additive model (GAM) fitted to the data, with the grey shaded area indicating the 95% confidence interval for the fitted curve. Log2CNR > 0 and log2CNR< 0 indicate regions with relative number gain and loss in BS, respectively. (b) The number and proportion of differential ATAC-seq peaks across the genome, on chr17, and on chr20, before and after applying CN normalization. (c) Averaged signal profiles of ATAC-seq and ChIP-seq in their corresponding peak sets. Supplementary Fig. 2. Impacts of copy number normalization on the differential analysis of ATAC-seq and ChIP-seq. (a) The proportion of more accessible, non-differential and less accessible ATAC-seq peaks in regions with different CNR without (top) and with (bottom) applying copy number (CN) normalization. Log2CNR> 0 and log2CNR < 0 indicate regions with relative number gain and loss in BS, respectively. (b) The trended biases of detected differential G-quadruplex formation activities driven by copy number differences without (left) and with (right) applying copy number normalization in ChIP-seq data. (c) Averaged signal profiles of ATAC-seq and ChIP-seq in their corresponding peak sets. Supplementary Fig. 3. Impacts of copy number normalization on differential anlaysis of Down syndrome ATAC-seq data. (a) MA plots of the genome-wide differential signals without (left) and with (right) applying copy number (CN) normalization. Differential peaks with adjusted P (p-adj) < 0.05 are depicted in blue while those with p-adj ≥ 0.05 are shown in grey. (b) A comparison of the differential signals without and with applying CN normalization for peaks on chromosome 21 (chr21) (blue) and other chromosomes (grey). (c) Differential signals for peaks on chr21 without (left) and with (right) CN normalization. (d) Differential signals for peaks on chromosome 17 (chr17) without (left) and with (right) CN normalization. Supplementary Fig. 4. No open chromatin region in the minimal critical region in Down syndrome. The top panel is genome browser screenshot of ATAC-seq tracks in genes near the Down syndrome critical region. Bottom panel is the zoom-in view of this regions with no ATAC-seq peak. Arrows on the gene models indicate the transcription direction. Supplementary Fig. 5. Example loci on chromosome 21 with dosage and compensatory effects in Down syndrome. (a) The genome browser screenshot of ATAC-seq tracks in interferon receptor gene clusters. Most of them showed increased total chromatin accessibility (shaded in dark orange) and one showed compensatory effects (shaded in blue). Differentially accessible (DA) regions depicted in red and blue bars represent regions with significantly increased and decreased chromatin accessibility, respectively. Arrows on the gene models indicate the transcription direction. (b) The genome browser screenshot of ATAC-seq tracks near the transcription start site of S100B gene. Both ATAC-seq peaks showed increased chromatin accessibility in total signal and averaged signal on each chromosome 21. Red bars indicate regions with significantly increased chromatin accessibility in Down syndrome. Arrows on the gene models indicate the transcription direction. Supplementary Fig. 6. Copy number ratio characterized by various copy number variation detection tools. Supplementary Fig. 7. Copy number ratio and distribution of differential signals identified by different pipelines on chr17 and chr20. Peaks with adjusted P (p-adj) < 0.05 are depicted in blue or red while those with p-adj ≥ 0.05 are shown in grey. Supplementary Fig. 8. Significantly differential peaks detected by HMCan-diff. (a) Example output from HMCan-diff taken from its official website. sb) Volcano plot of significantly different signals.