human gene annotation problems

[Mark Samuels]

I hope this has not been covered recently. When I launch IGV from the Broad site (v2.3, 750 Meg RAM), certain genes appear to be incorrectly annotated i.e. reading frames are wrong, in some cases ORFs don't even start with methionine and have multiple in-frame stop codons. I have not even loaded my sample bam files yet, this is still at the level of the basic annotation. I see this with both hg19 and b37 genomes loaded. Oddly, when I download the software (v.2.3.26)and install on my local computer, the same genes look fine in hg19 and mostly in b37. this is with the low memory version of IGV, but I don't think it's a memory problem as it is running with low memory allocation in the local installation too. I prefer not to use it locally if possible. Is there an incompatibility between the server-launchable 2.3 version and the current genome assemblies/annotations?

Mark Samuels

[Jim Robinson]

Hi Mark,

Could you give us a specific gene, preferably with a zoomed in screenshot so we can see the sequence and chromosome location?   There are annotations that do not start with "M" and have many stop codons, we would need specific examples to compare with the same annotation at UCSC to investigate futher.

Thanks

Jim

[Mark Samuels]

Hmm, I tried to upload screenshots to the forum site, it seemed like one could do that but it did not seem to work. Will email to you directly tomorrow. But these are well-annotated genes with human genetics, well-verified open reading frames etc. It seems to be specific to the version of IGV, not to the version of the genome, which is unobvious.

Mark

--
 
---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/972c4562-fd5b-425f-a885-94b8df579e70%40googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

[Jim Robinson]

You can email a screenshot to igv-...@broadinstitute.org.
Alternatively just name one (gene) and post the coordinates here.
We'll look into it with the latest version of IGV, and also verify it
with UCSC. We do not create annotation here, for human genomes we use
files from UCSC.

Jim

[Mark Samuels]

Attached are screenshots of the amino termini (not 5' ends of transcripts) of various genes chosen more or less randomly. In all cases, the reading frame is correctly interpreted using the downloaded IGV v2.3.26 (see the UCSC gene annotation screenshots at the end of the doc). However several of these genes are incorrectly translated using IGV v2.3 launched directly from the Broad site. In some cases they are wrong with both hg19 or (b37), in others the two genomes give different results. ATN1 is particularly odd as it seems to have no genomic sequence, even though the exons are aligned! 

I imagine that there are some mismatches between different versions of IGV and the underlying gene annotations. We had a problem with NextGene for a while, which they traced to the fact that inside the Refseq database some transcripts were defined as starting at position 0, and others at position 1 - this had nothing to do with the genomic nucleotide annotation but caused some genes to be out of frame when automatically translated, unless corrected for. Just to show that these things happen. But since many many people use IGV launched directly from your site, it would be good to rectify or at least alert to these issues. Feel free to call me at my office if you need more info. I should note that when I load my own aligned sample BAM files, more strange things happen again only with the IGV version launched from the site (so far).

thanks!

Mark Samuels

--

--- You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.

To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/52F038F8.2070405%40broadinstitute.org.

[Jim Robinson]

Hi Mark,

Thanks for the screenshots. Unfortunately they are not zoomed in enough
for me to read the sequence you are seeing, but that perhaps doesn't
matter. All versions of IGV use the same sequence and annotations,
regardless of where they are launched from. I do not understand what
version of IGV you are referring to when you say "IGV v2.3". The
website currently hosts version 2.3.26 (30) with a build date of
1/21/2014. I just launched this myself to verify it. The LDLR gene is
translated correctly (see screenshot). Perhaps you are having some
issue with caching that is preventing updates.

Jim

[Mark Samuels]

Hmm, you should be able to zoom in on the images I pasted into the Word doc, which would show nucleotide resolution.I can see in your screenshot that things are ok.

Anyway, when I go to the main Broad site and follow links to the IGV launch page, this is what I see (and others using it here as well, and also on my home computer). It says v2.3 at the top of the page, above the autolaunch buttons, and when launched has those gene annotation problems. If it is a cache problem, it is odd as this has been a problem since I first started trying to use IGV about two weeks ago (previously used NextGene but now am trying to visualize bam files created by collaborators, which is when the troubles commenced).

I guess this has to do with the autolaunch buttons, but as I say they seem to force me to that older version. I clear my browser cache regularly. Is there a java cache that needs to be cleared also, although not clear why that would affect what I see on your launch page.

Mark

--

--- You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/52F13F31.5010104%40broadinstitute.org.

[Jim Robinson]

There is a java cache, the details on how to clear it vary by platform
and OS version. A google search should turn up instructions. v2.3 is a
generic name, to get the exact version number do "help about".

The screenshots are not at high enough resolution to read the sequence,
one or two more zoom clicks would get them there. What I suspect is
happening is some network problem that only rears its head when you
start from Java Web Start. This could, for example, cause range-byte
headers to be stripped from the http requests for sequence => wrong
sequence => wrong protein translation. So I suggest you not use the
java web start buttons. We get ~10,000 hits on those links a week and
problems are rare, but they do occur.

Jim

[Mark Samuels]

Hmm, do you mean simply not to use the IGV launch buttons on your web site, but only use the version I have downloaded and installed locally? I can do, although it means some additional tweaking on my computer as the .bat command doesn't work due to my Java install not being clearly indicated in the Path environment variable apparently. My IT guy can help with that, I see instructions online but am hesitant to manually edit the Path myself, one typo and the computer is down. Odd that these things happen. Also for example another guy here at our hospital who is also launching IGV from your web site, noticed some genes were "funny" but simply ignored it. When looking at short read alignments to verify called variants, sometimes it's enough just to see the variant in the reads, not paying any attention to the gene annotation itself. But I like everything to be working whenever possible. Thanks for your assistance. It might be worth puting a general caution on your site that the auto launch may cause this problem for some users, so they can be alert to it and use a local version instead if necessary.

Regards,

Mark

--

--- You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/52F16395.5000808%40broadinstitute.org.

[James Robinson]

 Yes, that's what I'm suggesting.  We can't reproduce your problem here, but since a local install is working for you we can conclude that JWS is not.  We are testing a windows installer that will fix the path problem for you, watch for an announcement here.  

Jim

To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CAGNqJjiKb9Xv5B2mQhPBWcDNta2hxu3GwoVMMtej5mKp7Nmkmg%40mail.gmail.com.

[Mark Samuels]

Great, thanks James. Our bioinformatics analysis team (which is across town not on the hospital site) set up a procedure to launch IGV from the Broad site via our big CalculQuebec server cluster, which runs GATK and stores all our bam/bai/etc output files also. That way it is easy to tell IGV where the bam files are. (parenthetically, our hospital has a slow network connection to the big cluster across campus, so it would take days or weeks to physically transfer all the bam files, or else go running around town with a portable terabyte drive - and I've already had one crash that way). With a locally installed version of IGV, I can still get to my bam files, but telling IGV where to find them is a bit kludgier. Our bioinformics group runs IGV differently as well, so they also do not see the gene translation frame problems that we see here at the hospital. I have tried to reproduce the problem from my home laptop to see if it really is a Windows thing, but for unknown reasons I can't launch IGV at all from the Broad page, it starts and then crashes. I can run the locally downloaded version so it's not a java incompatibility. Anyway, at least I do have a workaround. The infrastructure support required for genome-wide sequencing is indeed not trivial. Running NextGene locally worked ok for me in the past, but it doesn't provide critical info like 1000 Genome allele frequencies, large exome datasets for filtering artifacts, SIFT, Polyphen, GenomeTrax, etc etc. I still like Nextgene as a teaching tool, as it gives students a chance to run their own analyses and get close to the raw data. But for state of the art genetic research it lacks capabilities.

Best,

Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CACOP%2Bptx-HixXpXj50e7iqewqx-%3DVBx%3Dk8NeaFyZChGAD7QXMQ%40mail.gmail.com.

[Jim Robinson]

Hi Mark,

If you could have someone from your core or bioinformatics group contact
us we can work with them and see if we can improve the IGV launch
procedure. Its difficult to give advice without details on what is
being done, and your network environment. As a workaround you can use
the IGV as they set it up to find the BAMs, save the session file,
then open your locally installed IGV and open the session file.

Also, you have not confirmed with us the version of IGV that is
displayed when you see the wrong annotation. The only way to do that is
to use "Help > About". I don't think this is related to IGV version,
but it is easy to check.

Jim

[Mark Samuels]

Jim, as I mentioned the bioinformatics core group that runs our basic pipeline is across down on a different network and using unix not Windows, so they don't see our problems with the web launch version. It would help if I could do the web launch from my home computer since that has nothing to do with our university or hospital network, firewall etc. That launch stalls partway through, I suspect it has trouble finding my java again with the path variable, but I am pasting below the various error messages that pop up. Let me know if you can figure out the issue so I can run the web launch version at home to see how the genes look from here.

Thanks!

Mark

 

Error: Found unsigned entry in resource: (http://igvdata.broadinstitute.org/app/current/igv.jar, 2.3.26)

Launch file window

<jnlp spec="6.0+" codebase="http://igvdata.broadinstitute.org/app/current" href="http://www.broadinstitute.org/igv/projects/current/igv.jnlp">

  <information>

    <title>IGV 2.3</title>

    <vendor>The Broad Institute</vendor>

    <homepage href="http://www.broadinstitute.org/igv"/>

    <description>IGV Software</description>

    <description kind="short">IGV</description>

    <icon href="IGV_64.gif"/>

    <icon kind="splash" href="IGV_64.gif"/>

    <offline-allowed/>

    <shortcut online="true">

      <desktop/>

      <menu submenu="IGV"/>

    </shortcut>

  </information>

  <security>

    <all-permissions/>

  </security>

  <update check="background"/>

  <resources>

    <java version="1.6+" initial-heap-size="256m" max-heap-size="750m"/>

    <jar href="igv.jar" download="eager" main="true" version="2.3.26"/>

    <jar href="batik-codec.jar" download="eager" version="1.7"/>

    <jar href="goby-io-igv.jar" download="lazy" version="1.0"/>

    <property name="jnlp.versionEnabled" value="true"/>

    <property name="java.net.preferIPv4Stack" value="true"/>

    <property name="apple.laf.useScreenMenuBar" value="true"/>

    <property name="com.apple.mrj.application.growbox.intrudes" value="false"/>

    <property name="com.apple.mrj.application.live-resize" value="true"/>

    <property name="com.apple.macos.smallTabs" value="true"/>

    <property name="http.agent" value="IGV"/>

    <property name="development" value="false"/>

  </resources>

  <application-desc main-class="org.broad.igv.ui.Main"/>

</jnlp>

Exception

com.sun.deploy.net.JARSigningException: Found unsigned entry in resource: (http://igvdata.broadinstitute.org/app/current/igv.jar, 2.3.26)

                at com.sun.javaws.security.SigningInfo.getCommonCodeSignersForJar(Unknown Source)

                at com.sun.javaws.security.SigningInfo.check(Unknown Source)

                at com.sun.javaws.security.JNLPSignedResourcesHelper.checkSignedResourcesHelper(Unknown Source)

                at com.sun.javaws.security.JNLPSignedResourcesHelper.checkSignedResources(Unknown Source)

                at com.sun.javaws.Launcher.prepareResources(Unknown Source)

                at com.sun.javaws.Launcher.prepareAllResources(Unknown Source)

                at com.sun.javaws.Launcher.prepareToLaunch(Unknown Source)

                at com.sun.javaws.Launcher.prepareToLaunch(Unknown Source)

                at com.sun.javaws.Launcher.launch(Unknown Source)

                at com.sun.javaws.Main.launchApp(Unknown Source)

                at com.sun.javaws.Main.continueInSecureThread(Unknown Source)

                at com.sun.javaws.Main.access$000(Unknown Source)

                at com.sun.javaws.Main$1.run(Unknown Source)

                at java.lang.Thread.run(Unknown Source)

--

--- You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/52F22251.5010706%40broadinstitute.org.

[Jim Robinson]

Mark, be patient for a day or two and we will have an installer for you
to test which should solve your Java problems. We will announce it here.

Jim

[Mark Samuels]

Great thanks Jim.

Mark

--

--- You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/52F27FA2.4000206%40broadinstitute.org.

[Mark Samuels]

Jim, I was finally able to get IGV running on a workstation with a lot of RAM. It did not matter whether I launched directly from your site with a jnlp or downloaded the software and ran locally, some genes were still problematic with some versions of the genome. So presumably this is a problem with our hospital network. Not sure if we can fix, certainly beyond my expertise. Question, while the software is making its calls to your database, is it possible to log what it is doing in plain text, or is it all just 0s and 1s? Would an IT professional on our site be able to figure out the problem, i.e.?

On a minor point, if I manually edit the downloaded bat file to change the memory allocation to java, is it just a question of changing the number, or do I need to do something special to make sure it's in 64-bit mode? Or is that just a function of the java version I have installed?

Thanks,

Mark

--

--- You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/52F27FA2.4000206%40broadinstitute.org.

[Mark Samuels]

Hey Jim, I think we have solved our problem, at least in one scenario. It was buried somewhat deep in IGV, but a new colleague who had done a lot with this at Baylor knew where to look. In the Preferences box, it showed that we were taking both our genome annotation and the bam files via a local university server. I don't know whether the whole genome is actually on that server or whether it was then making the call to your database, but in any case when we manually edited to go directly to you for the genome annotation, the problem seems to have gone away, and the bam files we load from the university server (which is outside our hospital network and firewall, but still here in Montreal) seem to align correctly. I think to do this we may have to use the locally installed version versus launching from your site, but that is a minor issue. thanks again, will let you know if the problem recurs.

Best,

Mark

--

--- You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/vmBxwU2ZwEM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/52F27FA2.4000206%40broadinstitute.org.