Modern software for whole genome alignment visualization

Question

[predeus]

Hello all,

I was wondering if there are any modern tools for visualization of whole-genome alignment, with ability to add annotation, and maybe extract variants? I've tried Mauve and Harvest so far, but they seem pretty abandoned and work quite poorly.

Are there any modern projects you can recommend? Surely with more and more genomes becoming available, there should be some tools to compare them? I am looking for something tailored to bacterial comparison, ideally, but any leads are welcome.

[mrals89]

I think there's a separation of concerns issue here OP. Genome browsers are typically responsible for loading BAM, fasta, and BED annotations, maybe vcfs. They are a data exploration application.

Ability to extract variants... you mean predict from the BAM files? Or you mean export the loaded vcfs? If you're talking about the former, as developers, we should keep the mathematics/algorithms separate from visualization applications such that we can iterate rapidly on both UI and server-side methods, rather than rely on one large monolithic application for everything. Does this make sense?

[predeus]

Maybe I did not make myself sufficiently clear. I have no problem choosing a classical genome browser (with one reference and one annotation) to view and analyze coverage, annotation, etc. Variant/coverage analysis for typical use cases (short reads aligned to a reference) are no problem as well.

I do, however, have a problem with similar software for comparison of 2 or more genomes. Basically, I was looking for more modern and more feature-rich versions of Mauve or ACT.

[mrals89]

Compare in what way, sir? Are you looking at rearrangements? CNVs? Or are you interested in homology and sequence features?

[predeus]

SNPs and indels with annotations (silent/missense/nonsense, frameshift/in-frame), for example.

[am2622]

Did you find a good browser to visualize 2 or more genomes + other tracks?

[predeus]

Nothing outside what was mentioned here, no.

This tool was useful: https://dnanexus.github.io/dot/

it uses Nucmer alignments and creates a pretty useful visualization.

[gina]

Did you reach any conclusions about a supported tool to do dynamic comparisons of multiple genomes? I am interested in genomes around 1.5 Gbases.

[predeus]

Hello Gina,

I came to conclusion that Dot tool mentioned above is very useful for a large-scale overview. Also, ACT, ancient as it is, worked quite well with the bacterial genomes I had to compare with annotation. But that might be too much for 1.5 Gb genomes. That's about it.

[rioualen]

Hi, any update about this? I'm also looking for a sorta genome browser to visualize whole-genome alignments for bacterial genomes, but I can't seem to make any of the tools mentioned here work -- many of which seem not to be maintained anymore...

[predeus]

Pretty much nothing outside things mentioned above. Dot is very good for large-scale comparison, and Artemis Comparison Tool (ACT) works nicely for detailed comparison using GFF annotation.

[lieven.sterck]

for a genome browser to visualise genome alignments, you can perhaps have a look at GenomeView. It's capable of visualising MAF alignments as tracks.

Answer 1

[lieven.sterck]

someone recently pointed me to this one: D-Genies . Haven't tried it myself though.

[predeus]

ha, that is a very creative use of minimap2 :)

at least there's a js visualization, worth a shot. Thank you.

Answer 2

[Alex Reynolds]

There's HiGlass, which is more for Hi-C datasets at this point, but a dotplot is effectively a heatmap and so may be usable for what you're doing.

It uses a "Google Maps"-like approach of scaling datasets at different resolutions, and so zooming in and out at different scales happens very quickly:

• http://higlass.io/
• https://github.com/hms-dbmi/higlass/wiki

Answer 3

[WouterDeCoster]

That sounds like a job for mummer http://mummer.sourceforge.net/

[predeus]

Last update on the code is in 2011, publications are from 2000-2004. Is it really still relevant? Visualizations look pretty horrible too (by modern standards of course).

[h.mon]

MUMmer 4 is in the works:

https://github.com/mummer4/mummer

[predeus]

Ah, thank you. This is very neat.

Answer 4

[tianleivv]

Have you tried Circos (http://circos.ca/)?

[predeus]

yes, of course. I think I should have specified that I was looking for a dynamic visualization - sort of like genome browser, but for 2 (or possibly more) genomes.

Answer 5

[Hussain Ather]

Mauve works too. http://darlinglab.org/mauve/mauve.html

[predeus]

yes, I did mention it in the original post... it's very poorly maintained, it seems. And since it's a Java jar, if it fails, there's no real way to debug it - you just have to accept it.

Answer 6

[h.mon]

I don't know of any software that meet all your needs, but you may try anvi'o and Artemis.

I've read somewhere (don't have the link) the The NCBI Genome Workbench may do what you want. I've never used it, though.

edit: web genome browsers (UCSC, GBrowse, JBrowse) are quite powerful, extensible and flexible, but probably not easy to set up.

[predeus]

Artemis Comparison Tool (ACT) is the tool related to Artemis, that does multi-genome visualization. It's very ancient and bad though, just painfully so.

Thank you for he anvi'o link, I will check it out though.

Answer 7

[AK]

CrossBrowse looks nice (biorxiv), although I failed to render gene models for my data... But the author is active at this moment (response to issue), perhaps we can all make it better. :-)

If you aim for visualizing variations (but from read alignment not genome alignment which you're mentioning), Sequence Tube Maps (interactive) might be a good idea:

Hope these help.