Cis and trans per sequence

Identical to its reference in the validationCis and trans sums as cooltools’ expected_cis and expected_trans make them (Open2C 2024). doi:10.1371/journal.pcbi.1012067

In the app

Analysis panel, under 2D tasks: Cis and trans per sequence

Command

chromosomix cistrans

Reads

A contact map

Also known as

interchromosomal, trans ratio, karyotype, cis/trans

How much of each chromosome’s contact stays on it (cis) and how much reaches the others (trans), and which pairs of sequences touch more than their depth predicts: whether a library worked, and whether the karyotype is what the assembly says. A pair well above the rest is a translocation, a fusion, a centromere cluster or a Rabl configuration, depending on the organism.

Running it

Open a contact map, choose Cis and trans per sequence in the 2D tasks card of the Analysis panel, untick any sequence to leave out of the results, and press run. The drawer shows the tables, the bars and the square below.

The same run from a shell:

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

Parameters

off-diagonal value decides what the pairs table and the square show. against indifference, the default, is the contacts between two sequences over what their own trans contacts predict: 1 is what any two sequences of that depth would give, and a fusion is far above it. raw contacts is the count, and per bin pair the count divided by the two sequences’ bin counts. least contact worth listing keeps a pair with a few stray reads out of the top of the table, and bars per sequence draws the cis fraction, the trans fraction, or both stacked.

Parameter	Takes	Default	What it does
sequences--sequences <name,…>	sequence names, comma-separated	Every sequence of the file when left out.	—
bars per sequence--bars <value>	cis, trans or both	cis	Fractions of each sequence’s own contact, so a deep sequence and a shallow one compare. both — the two sum to one, so every bar is full and the split is the cis fraction.
off-diagonal value--normalise <value>	enrichment, raw or bins	enrichment	enrichment — contacts over what the two sequences’ own trans coverage predicts. 1 is what any two sequences of that depth would give; a fusion is far above it; raw — raw contacts; bins — divided by the two sequences’ bin counts.
least contact worth listing--minContacts <n>	a whole number, at least 0	10	Raw contacts, before any normalisation, which keeps one stray read out of the table’s top.
pairs to list--maxEdges <n>	a whole number, at least 10	500	—
draw the square--matrix / --no-matrix	on or off	on	Skipped above 512 sequences, where it is not a picture.
use balancing weights--no-balance	a switch	On wherever the file carries balancing weights; --no-balance reads the raw counts.	—

What it writes

• table: a table: TSV
• curve: a curve: TSV
• matrix: a matrix, a pileup or a heatmap: TSV

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

Reading the result

Per sequence

The table lists the sequences in the file’s order, each with its Bins, its Cis, Trans and Total contacts, its Cis fraction (the share of its contacts that stay on it), Trans / cis and Contacts per bin. The bars draw the cis fractions in the same order, from 0 to 100 %. A sequence with no contact has no fraction: a gap among the bars and a blank in the table, never 0.

Between sequences

The square has a row and a column per ticked sequence, against indifference coloured blue below 1 and red above it; each sequence against itself, orders of magnitude above the rest, is left out. A cell well above 1 is a pair of sequences that touch more than their depths predict.

The pairs table ranks the same pairs, strongest first: both sequences, their raw Contacts and, against indifference, the Expected count beside Observed / expected. It says how many pairs passed the threshold when there are more than it lists.

What to watch for

• A summary, not a caller. Pairs are ranked against a stated null with no significance attached: a pair that stands out is not proof of a translocation.
• Balanced where the file is, unless you pass --no-balance. The map report runs this task on the raw counts.
• The square stops at 512 sequences, and the run says so; the pairs table and the bars have no such limit.