Somatic CNV Detection by Single-Cell Whole-Genome Sequencing in Postmortem Human Brain

Single-cell somatic copy number variants in brain using different amplification methods and reference genomes

“…Indeed, others have reported no major GC effect on MDA bias 23 . Although previous reports had shown superiority of PicoPLEX or MALBAC over MDA for CNV calling 18–23,31,32 , there had been no prior comparison to dMDA. The previous study of dMDA, conducted in lymphoblasts, reported more even genome coverage compared to MDA performed in parallel but not in droplets, although MAD values were not reported 26 .…”

“…We detected CNVs in ~30% of cells in two MSA cases in affected brain regions 16 , although their relevance remains unclear. We note several previous comparisons of WGA methods, with the broad consensus being that MDA is not suitable for calling CNVs by read depth, as there is bias due to over-amplification of certain regions, although it is more accurate at the single-base level [17][18][19][20][21][22][23] . There have been several attempts to reduce MDA amplification bias by performing reactions in nanoliter-scale volumes 10,24,25 .…”

“…We isolated single-nuclei using the CellRaft system (Cell Microsystems) mounted on a Nikon Eclipse TE300 inverted microscope coupled to a CCD camera (KERN optics), as described in detail elsewhere 16,18,46 . Briefly, we prepared nuclear fractions from 30-50 mg frozen tissues or cell pellets, and counterstained nuclei with 1 μg/ml DAPI for 20 min on ice.…”

“…The sequencing data were then sorted using Samtools-v.1.14 (RRID:SCR_002105; http://www.htslib.org) 49 and duplicates were marked and coverage metrics collected using Picard-v2. 18.4 (RRID:SCR_006525; https://broadinstitute.github.io/picard) 50 . BAM files were converted to bed files using Bedtools v2.…”

“…The gene lists were submitted to PANTHER (RRID:SCR_004869; http://www.pantherdb.org) 36 to identify statistical overrepresentation in any GO categories. The remaining analyses were conducted as described 16,18 .…”