Gabe Rudy's profile - activity

VCF files can be plotted in GenomeBrowse just like any other source. Plot > New Source > Browse, then select the parent folder and select the file.

One shortcut: you can also drag and drop VCF (and BAM) files directly into the "plot area" and a track will be added displaying the content of that file.

Hi there,

GenomeBrowse does not have any Copy Number analysis built-in, but the ploidy of the X chromosome is most likely visible at the right zoom level.

In general, you can zoom out and see the large-scale coverage depth at a chromosome level in GeomeBrowser. While the individual coverage will be noisy in regions, a general sense of the average depth between chromosome 22 and X would indicate if you have one, two or more copies of X.

Please be aware that GenomeBrowse is for education and academic use only and should not be considered diagnostic software.

Hi Jiwen, We see all our data servers and login server as being up. Let me know if your issue is persistent and what location you are in.

Thanks for the question. We don't have a way to change the default filter options on BAM plots currently, but will capture this as an area of improvement.

At wider zoom levels, GenomeBrowse needs an extra "precomputed" coverage file to be available for a BAM file to display coverage. A little "play" button should be in the top-left of a plot to start the computation to compute and write the .bam.covtsf file alongside the BAM file to display this coverage.

GenomeBrowse does not currently have a Sashimi like plot, but the next version will support viewing soft-clipped reads.

We support curating custom genomes that are not shipped with GenomeBrowse.

See this blog post for more details:

https://blog.goldenhelix.com/a-simple...

Hi there,

I might need a screenshot to provide a more concrete answer.

Depending on the zoom level, the height of the reads may start to become so small that the color appears darker. You can adjust the zoom in the vertical direction directly by togging the little zoom mode button in the top-right of the plot that has a letter like "A", "F", "H" or "M" on it.

When the view zooms out to show the precomputed coverage, the Pile-Up plot will show per-base mismatch data as black so it is easy to identify areas of mutation.

The BAM plots have two portions. The "Coverage" and "Pile-up". The pile-up will be the read fragments stacked efficiently for your inspection. As you zoom, you can see individual sequence details of your aligned reads.

Hi Signe,

The coverage track is gray when the correct reference sequence has been detected and so we can evaluate the match of the BAM letters to the reference (gray meaning matching bases, colored when mismatching).

First thing to check is that you have the appropriate reference sequences downloaded. You may have deleted the local reference sequence or you may be looking at BAMs with new assembly builds.

Hi Whit,

We have plans to make upgrades that will enable us to support reading CRAM files directly, but at the moment GenomeBrowse does not read them. CRAM is a complicated format and there is really only one good reference implementation of the code to read the file, so we have to make some infrastructure upgrades to get to the point where we can use that library directly.

Gabe

Hi Jonathan,

Currently GenomeBrowse will always display a heterogeneous genotype (regardless of phase) in the orientation of the reference on the top and the alternate on the bottom.

I see the appeal of being able to visualize the phase orientation, and it would probably warrant a slightly different display mode. But there is no current option to take the phase information into account.

Thanks, Gabe

I believe that issues is simply that your BAM file is in fact truncated (corrupted).

BAM files are essentially a specific type of compressed gz file with a series of compressed chunks. When reading a file for indexing, we are reading the entire contents and following the internal structure. This error says a given chunk expected more bytes than the file contained, usually indicating that the BAM file was copied from some other location but not in its entirety.

Without a valid BAM file, we can't build an index and display it.

It may be that although your files have a BAM extension, we are unable to open them if they are not readable as BAM files. The Golden Helix icon is just the GenomeBrowse file association, but does not grantee that the files are readable.

A couple of options are availabe.

First, you can drag and drop one or more BAM file into an open GenomeBrowse window.

Second, from the Plot dialog, try going to Browse and then selecting the folder containing the BAM files. You should be able to select multiple items from there.

when you convert your input data to an annotation track, be sure to specify the field you want to use to color your track as a "Categorical". This allows it to be selected in the style tab and each enumerated value can have a custom color mapping!

No, GenomeBrowse does not have an output that represents the per-base pileup information of a whole region.

The samtools mpileup tool may provide the type of output you are looking for

Hi there Frederick,

The "0/1 Genotypes" and Samples is the new names of the fields.

I just tried the latest GenomeBrowse and my own 23andMe data. The convert wizard showed no progress during the convert, but CPU was busy and it did complete. One thing to keep in mind is it does require that you have downloaded through the Data Source Library a copy of dbSNP and the the Reference Genome as we use those to get from the RSID fields in the input file to usable genotypes.

We also are not confidendent in the representation of insertions/deletions that 23andMe encodes, and other import tools ignore those, so best to ignore those.

Good luck! Gabe

Hi Tibor, It is possible, and the tricky thing with Linux is it is very hard for me to help reproduce / debug this. You may want to try updating to 18.04 (which the RHEL build works in our own testing). If that does not work, email me at my last name at goldenhelix.com

Hi Tibor, We are looking at bundling changes to be more compatible with more versions of Linux (binary compatibility is like walking a tight-rope). One thing you can try is moving the bundled `lib/libstdc++.so.6` out of the way. That may allow GB to run with the RHEL binaries.

HI Tibor,

It's unusual that this happened all of a sudden. There could be some issue with a file that the Plot dialog is trying to scan.

You can also try to install the RHEL version of GenomeBrowse, which we are considering making the default / only linux binary as it has been shown to be compatible with all linux distros:

http://www.goldenhelix.com/download/G...

Gabe

Hi Erin,

Please contact HGMD in terms of confirming if your licensed copy is available to integrate with any third party analysis solution.

Golden Helix has convert capabilities for standard genomic file formats such as BED/VCF for content that you own or have permission to use without restrictions inside third party tools.

Thanks, Jami...

Due to increased security requirements for connection to our login and upgrade server, existing GenomeBrowse installations (version 2.0.1 or newer) will fail to login with a red error message: “SSH handshake failed”. Unfortunately, for the same reason you will not receive the update notification for the recently released GB 3.0.0 that resolves this issue with an updated configuration.

You may find that GenomeBrowse continues to work for you if you had previously logged in and don’t log out, but please take the time to update to GenomeBrowse 3.0.0 to prevent any inconvenience and future interruptions.

Download and install GenomeBrowse 3.0 here: http://goldenhelix.com/products/Genom...

Thanks, Gabe

When logging in, I get a "SSH handshake failed" message!

Hey Michael,

Thanks for the feature suggestions. The "first in strand" coloring sounds interesting and I see why it would be a helpful way to visualize reads. I have created a ticket to capture that coloring mode. We are not actively adding features to GB currently, but may in the future have a GB feature push.

We looked into bigwig support and even had an experimental version working. Unfortunately it's a pretty funky binary file format from the UCSC guys and we had a number of issues with it. We will revisit this at some point as well.

WIG files are actually supported by GB. You need to "convert" them first. Just select them in the Convert wizard (from the data manager or Plot window).

Thanks, Gabe

Hey Michael,

One mode that might be useful is to change the stacking mode to Paired End. If there is a tandem duplication, you may see it in the boundaries where paired-end reads that span it don't find the standard distribution for distance between the paired end reads or if there is an inversion you don't see one of a forward and reverse strand being paired.

Click the little gear icon in the GB Pile-up plot and switch Stack mode to Paired End.

I hope that helps! Happy hunting! Gabe

Hey Magnus,

I would try refreshing / restarting GenomeBrowse.

When we push updates to the repo, it may be down for a couple seconds and you can try to load at those unfortunate times.

At the moment I don't see any connectivity issues in our logs or from our environment.

Also, make sure you have updated proxy settings if necessary in your environment!

Thanks, Gabe

Hey Ammar,

Our summary table here does not break out forward/reverse strand counts. Our default coloring scheme is based on forward/reverse mappings and there is also in the gear menu of the Pile-Up plot an option to "Split by Strand', which gives you a very nice visualization of strand-bias by placing all forward reads above and reverse reads below the axis.

Example stacking:

As an update to this, we now have native support for reading 23andMe files in GenomeBrowse!

Check out this YouTube video created by an avid GenomeBrowse user walking you through the process:

https://www.youtube.com/watch?v=ZHk4d...

Thanks, Gabe

Hey Beau,

I haven't been able to reproduce this on windows, so there might be something specific to the data or environment I'd like to track down.

But first, I'm going to send you an internal release for testing whether this has been fixed in our recent platform upgrade to Qt5 since our 2.0.7 release.

Thanks, Gabe

Hey Brian,

I think your accurate in pointing out we don't have great support for investigating structural varinats currently. It's definitely something I'd like to improve.

A couple of things we are considering (feel free to comment/expanding on these):

• Option to show soft and hard-clipped sequences when hovering over reads, allowing to see sequences at breakpoints.
• Option to color reads by mate-pair distances, with a special color for mates on other chromosomes
• Option to color reads with strand inversions from their mates
• Option to have hover-labels configured to any of these useful properties (mates position, mate distances, etc)

Is there anything else IGV does to support this we should consider?

Thanks, Gabe

Hey Eric,

Looks up to me. Could you let me know if there is a specific source that is having this issue?

Maybe post a screenshot.

Thanks, Gabe

Right, for our network BAM file reading, we assume the .bai file is located at file.bam.bai rather than file.bai.

I'm sorry that that assumption doesn't hold up in your situation.

For local BAM files, we can look at the file system and see if there is no file.bam.bai but there is a file.bai we can find the index file and move on.

We can not generate .bai files for remote sources as that requires reading the entire BAM file (which is as much work as downloading the file).

As an update

Hey An Hsu,

We don't allow plotting GTF files directly, but you should be able to convert them.

We decided not to display them in the Data Source Libraries view of plottable files in 2.0.2 since they are not plottable (that much has changed since 2.0.1)

But if you click on 'Convert', you should be able to Add your .gtf or .gtf.gz file and convert them into gene sources that you can use for plotting.

See our manual about converting GTF files.

Thanks, Gabe

Hi, I was trying to open BAM file in genomebrowser windows 64 bit version 2.0.1. But while doing that, the genomebrowser was shut down by itself every time. Does anyone has a solution for this?

Many thanks!

Hey An Hsu,

The BED format support in GenomeBrowse is focused around displaying "generic" intervals that have just a start, stop and potentially a label and color defined in the data.

If you actually have transcript data you want to display, I would suggest saving it out of UCSC in GTF format. The UCSC table browser should have a GTF export option.

Then our convert wizard will detect it as transcript data and you can visualize it as a Gene track.

Hope that helps, Gabe

Hey Enrique,

To answer your question from a Unix perspective: GenomeBrowse will see the file system like any other application and symbolic links are handled by the OS to look like the files they soft link to.

As long as the .bam and .bam.bai file are still side-by-side it shouldn't matter if they are being read through a symbolic link.

In fact, I use this quite a lot when testing etc.

Thanks, Gabe

Hey Anne-Katrin,

We do have plans to add a "Save as Image" capability in our next major release of GenomeBrowse (version 2.0). We are working on the feature now actually, but the roadmap for it's release depends on a couple other dependencies and probably won't be for 2-3 months.

Although we don't intend to ship GenomeBrowse with a full Python API like we do with SVS, we do plan on providing a "remote control" mechanism similar to IGV's URL based control mechanism. This means you could have links like:

http://localhost:60151/goto?locus=egfr

To jump the running instance of GB to that location (which could be coordinates, not just a gene name). We could similarly provide a way to save an image such as:

http://localhost:60151/saveimage?out=somefile.png&width=600&height=480

You could then programatically construct and call these URLs to take snapshots of a list of loci as you are suggesting.

If having a more supported path to this functionality or an expatiated roadmap is something you are interested in, let me know as we have provided services to other labs looking for custom GenomeBrowse features and support and I'd be happy to chat and discuss the scope of that.

Hey Oswaldo,

We do support itemRgb. We expect the RGB colors to be encoded like numbers 200,0,0 (no parens).

We also expect the itemRgb="on" key/val pair on the 'track' line of the file. For example, here is a couple lines of a file that we render with color:

track itemRgb="On"
chr7    54028   73584   uc003sii.2  0   -   54028   54028   255,0,0 .   AL137655
chr7    60328   61569   uc010krx.1  0   -   60328   60328   255,0,0 .   PDGFA
chr7    62967   63529   uc003sij.2  0   -   62967   63366   255,0,0 .   DQ576410
chr7    64068   64107   uc003sil.1  0   -   64068   64068   255,0,0 .   DQ584609
chr7    65159   65220   uc003sim.1  0   -   65159   65159   255,0,0 .   DQ600587
chr7    75460   116489  uc003sin.1  0   -   75460   75460   255,0,0 .   AL137655
chr7    244679  249951  uc003sio.1  0   +   244679  244679  0,255,0 .   AK024243
chr7    244826  249951  uc010kry.1  0   +   244826  244826  0,255,0 .   AK310146

I hope that helps.

-Gabe

Hey Bhakti,

Tandem Duplications may show up clearly in your alignments as either mismatches or large insertions, depending on your aligner's techniques.

Fusions may be a bit tricker to spot by eye, but fusion sites are often soft or hard clipped at the break point by the aligner for all spanning reads. It may look like a steep drop in coverage and the individual reads that you click on would show the clipping.

The evidence for fusion genes is almost always based on the paired-end alignments being one on each gene, and we don't provide a lot of support for visualizing that directly. If you go the gear menu in the pile-up plot and select to stack with paired-end matching, you will see pair-end reads that have their mate in a different chromosome or very far away in the same chromosome drawn with an arrow pointing towards their pair.

Another trick useful for this type of investigation is that you can open a second window on the same project from the Window menu. The second window can navigate to the second gene in the fusion event and by putting the windows side-by-side you can investigate the fusion.

But by the sound of your question, you might be searching for algorithms to detect tandem duplications and fusion events. If that's the case, I'm afraid GenomeBrowse won't be your final solution. GenomeBrowse is purely a visualization tool for existing data. If you have detected events, it's great at investigating their quality and genomic context, but it's not an analysis tool.

Thanks, Gabe

Hey Irina,

I'm glad you have found GenomeBrowse useful!

There is no normalization of the coverage data per say, but notice that the Y axis auto-scales to the extents of your data. So two samples side-by-side will show a similar profile of exon coverage even if one has 10x as much data.

If you want the Y-axis to be set to the same scale, you can select multiple plots and edit the Y axis limits in the plot's control panel.

On your second question, we have plans for a more extensive "save image" feature, but for now we suggest taking a screenshot and cropping it to your relevant plots.

Hey Rob,

I understand that this is a legitimate concern.

I can assure you that there no uploading of the contents of any of your data.

The purpose of the registration and account authentication is to allow for engagement through this community site and to have credentials for accessing secure cloud-based content (like our EA Pipelines and Illumina Basespace integration).

I have ideas about allowing for users to explicitly share projects and maybe even subsets of data as well through a cloud-based repository, but those of course would be things you opt in to very explicitly.

Thanks, Gabe