Per-cell contact QC

Not yet compared with its referenceThe near and mitotic contact shares of Nagano et al. 2017. doi:10.1038/nature23001

In the app

Analysis panel, under 2D tasks: Per-cell contact QC

Command

chromosomix cellqc

Reads

A contact map

Also known as

single cell, scool, cell cycle, quality

This task measures each cell of an .scool from its own contacts: its cis and trans shares, and its shares in the four bands of separation of Nagano et al. (2017), whose near band against its mitotic band draws the cell cycle as a horseshoe. Run it before you tick the cells to sum: a pseudo-bulk summed across phases smears the structure you opened the file to see.

Running it

Open an .scool, which opens on its Cells table, choose Per-cell contact QC in the 2D tasks card of the Analysis panel, leave cells on every cell in the file or pick the ticked ones only, and press run. The drawer shows a table with a row per cell and the phasing scatter. The shares also arrive as columns of the Cells table: sort by the near or the mitotic share, tick a run of cells, and sum them.

The same run from a shell:

chromosomix cellqc <map> [--<parameter> <value> …] --out <dir>

Parameters

The band edges are in base pairs, the paper’s numbers, and sized for mouse: on a smaller genome, set them to its scale.

Parameter	Takes	Default	What it does
General
cells--cells <value>	all or selected	all	all — what the choice is made from, so this is the answer that helps make it; selected — the ticked ones only.
use balancing weights--no-balance	a switch	On wherever the file carries balancing weights; --no-balance reads the raw counts.	—
Bands
near band, from--nearFrom <n>	a whole number, at least 0, in bp	23000	Below this a contact is mostly self-ligation and unligated ends rather than structure.
near band, to--nearTo <n>	a whole number, at least 1, in bp	2000000	—
mitotic band, to--mitoticTo <n>	a whole number, at least 1, in bp	12000000	The near band’s upper edge is this band’s lower one; beyond it is the far band.

What it writes

• table: a table: TSV
• curve: a curve: TSV

On the command line every file goes into --out as <file stem>.cellqc.<output>.<ext>, and each TSV opens with # lines naming the build, the file and every parameter that produced it.

Reading the result

The table has a row per cell: its name, whether it is Ticked, its Contacts and Stored pixels, its Cis and Trans shares, and its share in each band, headed by the band’s edges — < 23 kb, 23 kb–2 Mb, 2 Mb–12 Mb and > 12 Mb at the defaults. Every share is of all the cell’s contacts, trans included. A contact within one bin has a separation of 0 and counts in the first band. A cell with no contact has blank shares, not zeros.

In the phasing scatter each point is a cell, its near share across and its mitotic share up. Across a population the cells fall on a horseshoe, and where a cell lies along it is where it was in the cycle: just after mitosis a large mitotic band and little near contact, in G1 the reverse, and S and G2 between. The task assigns no phase: where to cut the horseshoe is yours, by ticking cells in the Cells table.

On an assembly whose longest sequence lies inside the near band, such as a yeast, the mitotic and far bands are empty and the scatter says nothing; the run’s note says so. Set the band edges to its scale.