IGV tracks fail to render even after zooming

Question

[mark.rose]

Hi All

I have a tool that performs alignment against a reference sequence and then creates an IGV session file based on that alignment and other features of the reference sequence.

<Global>
genome="dataset_18581.dat" locus="dataset_18581.dat" version="3"
<Resources>
<Resource name="726173-1-1 align" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//726173-1-1-1510602244-bowtie2.bam.sort.rmdup.bam.sort.bam"/>
<Resource name="726173-1-1 variants" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//726173-1-1-1510602244-bowtie2.bam.sort.rmdup.bam.sort.bam.vcf"/>
<Resource name="726173-1-1 zero coverage gaps" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//726173-1-1-1510602244-zero_coverage_gaps.bedgraph"/>
<Resource name="726173-1-1 below threshold coverage gaps" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//726173-1-1-1510602244-below_threshold_coverage_gaps.bedgraph"/>
<Resource name="726173-1-1 zero anchored coverage gaps" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//anchored-726173-1-1-1510602244-zero_coverage_gaps.bedgraph"/>
<Resource name="726173-1-1 below threshold anchored coverage gaps" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//anchored-726173-1-1-1510602244-below_threshold_coverage_gaps.bedgraph"/>
<Resource name="726173-1-1 soft clipping" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//COT102Frag1F4R1.txt-1510602244-soft_clipping.bedgraph"/>
<Resource name="726173-1-1 non-ref frequency" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//COT102Frag1F4R1.txt-1510602244-non-ref_freq.bedgraph"/>
<Resource name="726173-1-1 strand bias" path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//COT102Frag1F4R1.txt-1510602244-strand_bias.bedgraph"/>
<Resource name="726173-1-1 homopolymers " path="http://demeter//project/LEE2705/TEST/TK0000000/171013_M01610_0074_000000000-D30NY/726173-1-1/1510602244//COT102Frag1F4R1.txt-1510602244-homopolymers.bedgraph"/>
</Resources>
</Global>

I have used this several times and it has worked fine. However, I have a dataset now where several of these tracks are not rendered though they are uploaded (because they are present in the track window) and they are read by IGV (since the displayed data range is correct for each track that contains data (not all do)). For this dataset the reference is small (1426bp) and the reads are 250bp PE. The alignment (bam) and coverage tracks seem to render properly but tracks showing strand bias and homopolymers (which contain data) are empty. I can zoom in as far as possible at locations where features should appear but there is nothing. Likewise no choice of "windowing function" makes them appear either.

Any suggestions

Mark

[Devon Ryan]

Have you indexed the bedgraph files?

[mark.rose]

I presume you mean with tabix. No, but I have never done that in the past and never, in the many times I've done this, had any issue. Is there a reason why indexing would be needed in some cases but not others? This is a smaller reference than I have used in the past.

[Devon Ryan]

I'm not sure if tabix is needed or if IGV uses it's own thing, but either way. Often IGV runs into memory issues if you don't index things like this. If that's the cause then it's an easy enough fix.

[mark.rose]

Well I looked more online and found some people using tabix for indexing bed files for IGV and so that is what I did (after commenting the header)

sort -k1,1 -k2,2n test.strand_bias.bedgraph | bgzip > test.strand_bias.bedgraph.gz; tabix -s 1 -b 2 -e 3 test.strand_bias.bedgraph.gz

I then substituted the in path to this gzipped file in my session file and reloaded. The only difference is that the line which separates positive and negatives values (100 to -100) tuned blue but still no features.

[Devon Ryan]

Do you see any errors printed to the terminal? If you're not starting IGV from a terminal, do so so you can see these errors. At the end of the day that's likely the only way to diagnose this.

[mark.rose]

I am launching it on a windows machine with a igv.bat file. When I do so a terminal window as well as the viewer appears. There is no error message in the terminal just the usual startup info ![https://ibb.co/gij7nb][1]

Also I tried upgrading to version 2.4.4. No difference.

I then tried dropping the number of reads to align to my 1.4kb reference to 100 (250bp PE). Java has 2G of allotted memory to run IGV. I would think this would be way more than enough if there is some memory limitation. While I do now get some tracks with data showing up, they are ones reporting on aspects of coverage, which of course now is low with only 100 reads aligned. The homopolymer and strand bias track remain as before, blank.

Answer 1

[Jim Robinson]

Hi, is your chromosome name really COT102Frag1F4R1.txt?

Its possible there are no records in the bedgraph file that match the sequence names in your reference fasta. That's just a guess though.