SUPR
High percision reconstruction of single nucleus
Dnr:

NAISS 2025/22-124

Type:

NAISS Small Compute

Principal Investigator:

Jinxin Chen

Affiliation:

Karolinska Institutet

Start Date:

2025-01-24

End Date:

2025-08-01

Primary Classification:

10610: Bioinformatics and Computational Biology (Methods development to be 10203)

Webpage:

Allocation

Abstract

DNA is the fundamental genetic blueprint which every single cell of our body has and follows its command. While DNA is 2 meters long, the nucleus of the cell which holds them is 10 micrometers in diameter. How such a long strand of DNA is compacted inside the nucleus remains elusive. To study such organization of DNA and its functional impact on gene regulation, we combine Fluorescence in-situ hybridization (FISH) with single-cell Hi-C (scHi-C) to generate high-resolution 3D structure of a single cell, with which detailed structures in genome can be studied. In the study, we generate whole-genome FISH data with 900 probes at the resolution of 3 Mb and scHi-C from the exact same cells. The whole-genome FISH data will be severed as scaffold with the scHi-C contact data will be the filling that enables the high-resolution reconstruction. The raw data of sequencing reads are demultiplexed to fastq files using bcl2fastq (v2+). After it, the reads will be aligned to the GRCh38 human reference genome using bwa-mem (0.7.13-r1126) and hisat-3n (2.2.1-3n-0.03). Then, the reads are paired to contact with hickit (https://github.com/lh3/hickit) and contact files are generated. Rough 3D reconstructions are simulated with Dip-C (https://github.com/tanlongzhi/dip-c). Then the raw imaging files are deconvolved with Deconwolf (https://github.com/elgw/deconwolf), after which nuclei and dots inside the nuclei from the possessed imaging data are extracted by Dotter (https://github.com/elgw/dotter). Then, dots are overlapped over all the rounds with the calibration by fiducial marker and decoded. Lastly, we combine the FISH data and sequencing raw structure with modified nuc-dynamic simulation algorithms.