Announcement

Collapse
No announcement yet.

Irys View, SV detection, smap files

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Irys View, SV detection, smap files

    Hi,

    I seem to not be able to find out where I can get the SV detection file (smap) from, or how to generate it in the first place. The documention is extremely unspecific on this and I do not see Irys View options, not are the smap present in my workspace folder.
    (i was running cmap vs cmap comparisons in an separate workspace from the one where I did the initial assembly - the very un-organized way IV shows all the results in the "Comparison" etc boxes forced me to do this to not loose overview entirely).
    So how can I run a SV detection?

  • #2
    here the answer from Kees-Jan for anyone who might look for the same thing


    You can run the SV detect script from the scripts directory. It’s called runSV.py to do a SV detection. If you run this script with the –h parameter (python ./runSV.py –h) you will see all the available options and requirements. By selecting one assembly (cmap) as reference and one other assembly as individual (query, Assembly output)) you will obtain the SV’s between two individuals.
    You can visualize the resulting maps in irysview.

    Comment


    • #3
      ... rewritten post after I got logged off while writing the previous version of this entry.

      for the record and to report some odd things:
      I successfully ran the SV detection:

      python /home/estolle/scratch/irys/scripts/runSV.py \
      -t /scratch/irys/tools/RefAligner \
      -r /scratch/irys/SVdetect/Mir6B_LGs.cmap \
      -q /scratch/irys/SVdetect/in \
      -o /scratch/irys/SVdetect/SV_Mir6b_vs_Mir6BLGs \
      -p /scratch/irys/scripts/ \
      -a /scratch/irys/SVdetect/Mir6B_June2016B.xml \
      -T 40 \
      -j 10

      I found alot of things I expected it to find. However. Some SV's despite being large are missed. Also, I noticed some imprecisions in which labels are used to determine the extent of the SV. The lasso tool to measure sizes is very odd to use. It seems I need to estimate the reference-scaffold/cmap region I want to display (eg the SV) on the query strand straight underneath - not where the homologous labels are. While I can figure out some size differences (of the SV) even when neighboring labels/regions are also marked, this is pretty much useless in this form.
      The other odd thing is. When I displayed the "filtered" SVs bedfile, then their coordinated are way off and do not match a SV at all anymore. The unfiltered SV-bedfile ones seems to be correct. During the filtering most of the SVs I actually want to detect, are filtered out. even big/clear one. Its not quite clear to me how the SVs are filtered / against which criteria.

      Some examplepix attached

      Comment


      • #4

        Comment


        • #5

          Comment


          • #6
            Hi Eckart - frustrated? I went through that too.... Here is a custom solution that we use succesfully to extract all those visible SVs. This is a passage from our recent paper, follow the github link below to download an easy to use python script to extract SVs and export as bed file. Input is the xmap, q.cmap and r.cmap as well as a cutoff value for size. Lower cut off is 2500bp in Arabidopsis - experimentally verified. SVs <2.5kb are two close nicks in ref, that is computed as single nick in cmaps. Contact me directly if you have more questions [email protected]

            Florian


            from: Kawakatsu et al. 2016 Cell; http://dx.doi.org/10.1016/j.cell.2016.06.044
            To call SVs we used initially the BioNano own command-line script and compared this to a manual analysis in IrysView (BioNano Genomics). This however identified deviations between reference and query, and the BioNano own script only reported less than 25% of the manually identified SVs. We tested the result by assessing single molecule quality and coverage at potential SVs over 2.5 kb, and found all positive (similar to (Mak et al., 2016)). To call SVs we developed our own pipeline (available at https://github.com/RyanONeil/structome) using python scripts to parse the positions of all insertions, deletions and inversions from the .xmap and .cmap files. Insertions and deletions were computed by calculating the differences between aligned nick positions within the reference and the aligned maps. With resolution of only 1 kb we were careful in selecting a cut-off of 2.5 kb, as smaller differences in alignment resulted in an abundance of false positives. Inversions were detected as breaks in the alignment where the two contiguous segments originating from the same cmap mapped to neighboring regions of the reference but in opposite directions.

            Comment


            • #7
              Hi Florian

              Thats great. Thanks for the hint! I am doing a recheck of my SVs now with your scripts. Very helpful!
              cheers

              Comment

              Working...
              X