Missing arcs when plotting BEDPE file format

[eafyounian]

Hi! When plotting arcs for BEDPE file format, if the first coordinate comes after the second coordinate (i.e. start_x > start_y), the arc is not drawn. I have fixed this issue by swapping the first coordinate with the second coordinate (see figure where the first track after the gene track shows some features, the second track is how it looks after my fix, and the third track is before the fix where some arcs are missing).

Is there some other more elegant fixes?

By the way, I am using coolbox version 0.3.3. I quickly checked the source of the latest version (i.e. 0.3.8) but I did not immediately find any indication that this is addressed in the latest version (but of course I may have missed it 🙂).

Thank you in advance for your consideration and response! 🙂

[Nanguage]

I have reproduced this behavior, the Arc track fetched the non-upper-triangular record(start_x > start_y) but it's not drawn.

After I checked, this bug is caused by the following code:

CoolBox/coolbox/core/track/arcs/plot.py

Line 101 in 36a86b2

diameter = (end - start)

When drawing, the diameter of arc is determined by pos2 - pos1 and becomes negative when plot the non-upper-triangular record(start_x > start_y).

A more elegant solution is to add a judgment to fetch_plot_data that converts all records to the upper-triangular form(start_x < start_y). fetch_plot_data is this function:

CoolBox/coolbox/core/track/arcs/bedpe.py

Lines 56 to 78 in 36a86b2

	def fetch_plot_data(self, gr: GenomeRange, **kwargs) -> pd.DataFrame:
	df = self.fetch_data(gr, **kwargs)
	if len(df) == 0:
	return df
	style = self.properties['style']
	if style == self.STYLE_ARCS:
	pos_at = self.properties['pos']
	if pos_at == 'start':
	pos1, pos2 = df['start1'], df['start2']
	elif pos_at == 'end':
	pos1, pos2 = df['end1'], df['end2']
	else:
	pos1, pos2 = (df['end1'] + df['start1']) / 2, (df['end2'] + df['start2']) / 2
	return pd.DataFrame({'pos1': pos1, 'pos2': pos2, 'score': df['score']})
	elif style == self.STYLE_HICPEAKS:
	gr2 = kwargs.get('gr2')
	if gr2 and gr2 != gr:
	mask = (df['start2'] >= gr2.start) & (df['end2'] <= gr2.end)
	df = df[mask]
	return df
	else:
	raise ValueError("The supported style for bedpe data are ['arcs', 'hicpeaks']")