Seqanswers Leaderboard Ad

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • sterding
    Member
    • Sep 2010
    • 36

    gap threshold in joining the reads in cufflinks?

    I have a question to the de novo assembly using cufflinks: is there a threshold to join two non-overlapping reads?

    As you can see from the attached figure, the top track is for alignment (BAM file from Tophat for non-strand-specific RNAseq), and the lower track is assembly from cufflinks. I run this simply as

    cufflinks input.bam

    It's for sure that overlapped reads are joined together into one isoform (or transfrag). What's strange for me is, why some separated reads (with gap between) are also joined, but some are not. As you can see, there are 4 gaps in the top track, where the first and last gap are not joined, while the 2nd and 3rd ones are joined. I guess it's because of the smaller gap distance in 2nd and 3rd gap, but I cannot find such 'gap threshold' in the cufflinks options.

    I also tried the command with options:

    cufflinks --min-intron-length 100
    and
    cufflinks -F 0.0 --trim-3-dropoff-frac 0 --trim-3-avgcov-thresh 0 --min-frags-per-transfrag 1 --min-intron-length 100

    both don't change the result.

    I appreciated if any could explain this to me. Thanks a lot.
    Attached Files
    Last edited by sterding; 04-05-2012, 06:20 PM.
  • sterding
    Member
    • Sep 2010
    • 36

    #2
    OK. Without any reply yet, I dig into the source code of cufflinks by guessing something with threshold of 50. Yes, I did find some secret.

    Code:
    [sterding@hpcc01 cufflinks-1.3.0]$ grep -P "50\D" -I src/*
    src/bundles.cpp:				if (counter.total_reads < 100 || overhang_ratio >= 0.50)
    src/common.cpp:uint32_t min_intron_length = 50;
    [COLOR="Red"]src/common.cpp:int olap_radius = 50;[/COLOR]
    src/common.cpp:float high_phred_err_prob = 0.50; // about MAPQ = 3
    src/cufflinks.cpp:    fprintf(stderr, "  --min-intron-length          minimum intron size allowed in genome                 [ default:     50 ]\n");
    src/cufflinks.cpp:    fprintf(stderr, "  --overlap-radius             radius size to define overlapping reads               [ default:     50 ]\n");
    src/GArgs.cpp:      <string>{;|=} e.g. disable-test;PID=S= for --disable-test PID=50 (or --PID 50) S=3.5 etc.
    src/GArgs.h:       <string>{;|=} e.g. disable-test;PID=S= for --disable-test PID=50 (or --PID 50) S=3.5 etc.
    See the read part: overlap_radius = 50. That's what I guessed. And this parameter seems not changeable for the user:


    Code:
    [dongx@hpcc01 cufflinks-1.3.0]$ grep -P "radius" -I src/*
    src/assemble.cpp:		[COLOR="red"]fill_gaps[/COLOR](scaffolds, 2 * olap_radius);
    src/assemble.cpp:		[COLOR="red"]fill_gaps[/COLOR](new_scaffs, 2 * olap_radius);
    src/bundles.cpp:                        (!Scaffold::overlap_in_genome(L, R, olap_radius) ||
    src/bundles.cpp:		if (bh->ref_id() == bundle.ref_id() && bh->left() < bundle.right() + olap_radius)
    src/common.cpp:int olap_radius = 50;
    src/common.h:extern int olap_radius;
    src/cuffdiff.cpp:    olap_radius = 0;
    src/cufflinks.cpp:{"overlap-radius",       required_argument,	     0,			 OPT_OVERLAP_RADIUS},
    [COLOR="red"]src/cufflinks.cpp:    fprintf(stderr, "  --overlap-radius             radius size to define overlapping reads               [ default:     50 ]\n");[/COLOR]
    src/cufflinks.cpp:				olap_radius = parseInt(0, "--overlap-radius must be at least 0", print_usage);
    src/graph_optimize.cpp:        compressed.fill_gaps(2 * olap_radius);
    src/scaffold_graph.cpp:				hit_j->left() - hit_i->right() < olap_radius)
    src/scaffolds.cpp:								 int overlap_radius)
    src/scaffolds.cpp:	int ll = lhs.left() - overlap_radius;
    src/scaffolds.cpp:	int rr = rhs.right() + overlap_radius;
    src/scaffolds.cpp:	int lr = lhs.right() + overlap_radius;
    src/scaffolds.cpp:	int rl = rhs.left() - overlap_radius;
    src/scaffolds.cpp:		if (overlap_in_genome(lhs, rhs, olap_radius))
    src/scaffolds.cpp:		if (overlap_in_genome(rhs, lhs, olap_radius))
    src/scaffolds.h:								  int overlap_radius);
    I might re-compile the code and let the "--overlap-radius" reachable for the user. Or, It's really appreciated if cufflinks team can do this.

    Comment

    • CHLin
      Junior Member
      • May 2012
      • 3

      #3
      This might help...

      I think this option may help to change the parameter:
      --overlap-radius [#nt]
      If the gap between two units is larger than the #, they will be merged.

      Comment

      • sterding
        Member
        • Sep 2010
        • 36

        #4
        Originally posted by CHLin View Post
        I think this option may help to change the parameter:
        --overlap-radius [#nt]
        If the gap between two units is larger than the #, they will be merged.
        You are right; this option becomes user-settable until the recent release.

        Comment

        Latest Articles

        Collapse

        • seqadmin
          Pathogen Surveillance with Advanced Genomic Tools
          by seqadmin




          The COVID-19 pandemic highlighted the need for proactive pathogen surveillance systems. As ongoing threats like avian influenza and newly emerging infections continue to pose risks, researchers are working to improve how quickly and accurately pathogens can be identified and tracked. In a recent SEQanswers webinar, two experts discussed how next-generation sequencing (NGS) and machine learning are shaping efforts to monitor viral variation and trace the origins of infectious...
          03-24-2025, 11:48 AM
        • seqadmin
          New Genomics Tools and Methods Shared at AGBT 2025
          by seqadmin


          This year’s Advances in Genome Biology and Technology (AGBT) General Meeting commemorated the 25th anniversary of the event at its original venue on Marco Island, Florida. While this year’s event didn’t include high-profile musical performances, the industry announcements and cutting-edge research still drew the attention of leading scientists.

          The Headliner
          The biggest announcement was Roche stepping back into the sequencing platform market. In the years since...
          03-03-2025, 01:39 PM

        ad_right_rmr

        Collapse

        News

        Collapse

        Topics Statistics Last Post
        Started by seqadmin, 03-20-2025, 05:03 AM
        0 responses
        49 views
        0 reactions
        Last Post seqadmin  
        Started by seqadmin, 03-19-2025, 07:27 AM
        0 responses
        57 views
        0 reactions
        Last Post seqadmin  
        Started by seqadmin, 03-18-2025, 12:50 PM
        0 responses
        50 views
        0 reactions
        Last Post seqadmin  
        Started by seqadmin, 03-03-2025, 01:15 PM
        0 responses
        201 views
        0 reactions
        Last Post seqadmin  
        Working...