[Guidobot]

Hi Iain,

I sent you an email and the datasets. I also asked a question that I think would be good to answer here.

What does the new contig. Mismatch value report refer to?

If it is the total number of mismatches w.r.t the consensus/reference across all reads, is there a minimum number/percentage of reads required to denote a difference (i.e. probable SNP)? Does it also include mismatches for reads counted as features (i.e. with INDELs)?

Also the value appears to be rounded (%g) so it is hard to know the decimal point accuracy.

[imilne]

What does the new contig. Mismatch value report refer to?

It's a count of the number of read bases (across all reads) that don't match the consensus/reference at that base. It doesn't attempt to do anything clever (because we need to calculate the number very fast), so if a read's base doesn't match the consensus - no matter the reason - then it gets counted.

The decimal places may be rounded for display purposes in the table, but if you right-click it and copy its data to the clipboard, you'll get the full number.

Iain

[m_welkers]

Exporting reads intersecting column

Just want to say, I really like tablet and use it frequently! I do have some quick questions. I work with influenza datasets and when I load in my datasets to look for SNP's, I go to that position in Tablet and export all the reads that intersect that column. The reads are exported to a file in a .fas format and I noticed a few things. First off the "pos=" value does not match the same starting position displayed in Tablet but is off by 1. (So in Tablet start position of the read is 2155 and exported it says pos=2154. The CIGAR string is 54M so no padding or other issues with the string.)

Beyond that I'm curious what it does with insertions. It doesn't appear to display them in the read so it doesn't extract the sequence directly from the reads in the alignment file. How does it "make" the sequences then?

Ps. Keep up the good work!

[imilne]

First off the "pos=" value does not match the same starting position displayed in Tablet but is off by 1.

Yep, that one's a bug (and has already been fixed), so it'll be in the next release.

Beyond that I'm curious what it does with insertions.

What kind of? CIGAR ones?

Iain

[Guidobot]

What does count as a Feature in Tablet?

I originally thought this meant just CIGAR strings with inserts but I have examples with more and less reads with inserts than reported Features.

[JohnN]

Is it on the drawing board that Tablet would eventually export the consensus sequence?

John

[imilne]

Is it on the drawing board that Tablet would eventually export the consensus sequence?

It already can. Just right-click the consensus sequence and select a copy-to-clipboard option.

Iain

[JohnN]

Thanks... I looked and looked but obviously didn't look hard enough

[imilne]

Thanks... I looked and looked but obviously didn't look hard enough

Or we didn't put it in an obvious place. Where did you look (or expect to see it)?

Iain

[JohnN]

Or we didn't put it in an obvious place. Where did you look (or expect to see it)?

Iain

I thought I had looked in the documentation, but I was looking for a specific command which read "Copying consensus to a file" or something like that. After your "try right-clicking" post above, I looked in the documentation and found the following:

Right clicking with the mouse on many of the display components will open up additional menus showing options to change the display types, highlight regions of interest, copy data to the clipboard, jump to a read's pair if it is a paired read, etc.

You can't be any clearer than that. Now that I feel silly, this post may stop others making the mistake that I made, if they use "search".

Thanks again,
John

[imilne]

Here's a summary of the changes in the latest release (1.11.08.10):

- NEW: Added a new colour scheme for displaying Read Groups in SAM and BAM files.
- NEW: Added a new control panel tab for viewing and editing Read Group colours.
- NEW: GFF files now support drag and drop loading.
- NEW: GFF files can now be specified on the command line (or via web start).
- NEW: Reduced the initial detection time for BAM files, which should speed loading.
- NEW: Added a tablet.xml preferences variable which governs the number of insert events a CIGAR-I feature has to relate to before it is included.
- CHG: String values longer than 75 characters in the graphical tooltips are now be truncated.
- CHG: All searches are now case insensitive.
- CHG: The setting for using less-strict BAM validation checks is now on by default.
- BUG: Toggling Show Bases wasn't causing the screen to refresh.
- BUG: Fixed a crash that could occur when searching over consensus sequences.
- BUG: The recent-file tooltips on Linux and OS X weren’t working with paths beginning with "/".
- BUG: RegEx characters included in a BAM file’s name (eg ‘[]’) were stopping Tablet from locating the corresponding BAI file.

As usual, Tablet should auto-update the next time you run it, or you can grab the download manually from http://bioinf.scri.ac.uk/tablet

Iain

[AnthonyB]

Hi,
I have been experimenting with tablet to view de novo assemblies produced with MIRA (via conversion from maf to bam). I was wondering if their was a way to set the paired-end coloring to highlight regions of paired-end inconsistencies (too short, too long, map to another contig etc) rather than the standard read direction/unpaired/no mate scheme which is currently employed?
Thanks,
Anthony

[GStephen]

There's not currently a way to achieve what you are talking about with Tablet, but we're actively looking in to providing further paired-end visualizations. There may be some updates to this functionality soon.

Gordon

[maubp]

Hi,
I have been experimenting with tablet to view de novo assemblies produced with MIRA (via conversion from maf to bam). I was wondering if their was a way to set the paired-end coloring to highlight regions of paired-end inconsistencies (too short, too long, map to another contig etc) rather than the standard read direction/unpaired/no mate scheme which is currently employed?
Thanks,
Anthony

As Gordon says, Tablet doesn't (yet) have any length based paired end colouring. However, I think the standard paired colouring will highlight reads where the partner is either unmapped or mapped to a different contig.

Also, in order to know if a pair is too short or too long, the viewer will need to know the expected range. For SAM/BAM, this can be specified in the read group header (the @RG line), but currently the maf2sam.py converter does not do this. I assume that's how you're converting your MIRA file to BAM:

http://github.com/peterjc/maf2sam

Test data would help if you would like that added to the maf2sam.py script (although given recent discussion on the samtools-devel mailing list, native SAM output from MIRA might be happening before too long).

[AnthonyB]

As Gordon says, Tablet doesn't (yet) have any length based paired end colouring. However, I think the standard paired colouring will highlight reads where the partner is either unmapped or mapped to a different contig.

Also, in order to know if a pair is too short or too long, the viewer will need to know the expected range. For SAM/BAM, this can be specified in the read group header (the @RG line), but currently the maf2sam.py converter does not do this. I assume that's how you're converting your MIRA file to BAM:

http://github.com/peterjc/maf2sam

Test data would help if you would like that added to the maf2sam.py script (although given recent discussion on the samtools-devel mailing list, native SAM output from MIRA might be happening before too long).

Yes, I'm using maf2sam.py for the conversion and I had noticed that there was currently no implicit paired-end distance set in the SAM file (Tablet reports all the lengths as 0). I'd be happy to share some MIRA assemblies if it would help, but it seems that this type of shading is beyond Tablet's functionality (at least for the time being). Also, as far as I can see (and please let me know if there something I'm missing), Tablet doesn't differentiate in-contig with off-contig mapped pairs by color, only those with unmapped pairs.

Thanks also for the heads up about MIRA and SAM output this should be useful to try out if/when it can be implemented.
Anthony