HICCUPS FDR threshold

357 views
Skip to first unread message

Kate Cook

unread,
Mar 14, 2016, 8:41:23 PM3/14/16
to 3D Genomics
Hi,

When I run HICCUPS with a loose FDR threshold of 0.9, I get this loop reported:

chr1    x1    x2    chr2    y1    y2    color    observed    expectedBL    expectedDonut    expectedH    expectedV    fdrBL    fdrDonut    fdrH    fdrV    numCollapsed    centroid1    centroid2    radius
chr5    430000    440000    chr5    600000    610000    0,255,255    35.0    12.019885    11.546831    12.39225    10.399725    0.001752831    0.0017599224    0.09893618    0.0017469053    2    435000    600000    25000

But when I run it with a threshold of 0.1, that loop is absent, even though all four FDR values are <0.1.

Am I missing something about how that FDR threshold is used? I thought that all four (doughnut, BL, H, V) had to be below the threshold.

Thanks,
Kate

Muhammad Saad Shamim

unread,
Mar 14, 2016, 9:39:19 PM3/14/16
to Kate Cook, 3D Genomics
Hi Kate,

What were the exact parameters/flags for the HiCCUPS run?

--
You received this message because you are subscribed to the Google Groups "3D Genomics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/3d-genomics/1bba890d-1f2f-4820-be20-91dd7cd399a9%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Suhas Rao

unread,
Mar 14, 2016, 10:06:58 PM3/14/16
to 3D Genomics, kate...@uw.edu
Also, what list does this entry appear in? The final merged list (i.e. after post-processing) or the first list of enriched pixels?


On Monday, March 14, 2016 at 6:39:19 PM UTC-7, Muhammad Shamim wrote:
Hi Kate,

What were the exact parameters/flags for the HiCCUPS run?
On Mon, Mar 14, 2016 at 7:41 PM, Kate Cook <kate...@uw.edu> wrote:
Hi,

When I run HICCUPS with a loose FDR threshold of 0.9, I get this loop reported:

chr1    x1    x2    chr2    y1    y2    color    observed    expectedBL    expectedDonut    expectedH    expectedV    fdrBL    fdrDonut    fdrH    fdrV    numCollapsed    centroid1    centroid2    radius
chr5    430000    440000    chr5    600000    610000    0,255,255    35.0    12.019885    11.546831    12.39225    10.399725    0.001752831    0.0017599224    0.09893618    0.0017469053    2    435000    600000    25000

But when I run it with a threshold of 0.1, that loop is absent, even though all four FDR values are <0.1.

Am I missing something about how that FDR threshold is used? I thought that all four (doughnut, BL, H, V) had to be below the threshold.

Thanks,
Kate

--
You received this message because you are subscribed to the Google Groups "3D Genomics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics+unsubscribe@googlegroups.com.

Kate Cook

unread,
Mar 14, 2016, 10:55:06 PM3/14/16
to Suhas Rao, 3D Genomics
These are the commands I'm using:

./juicebox.sh hiccups -r 10000 -f 0.9 -p 1 -i 5 sporozoite.hic hiccups_loops.txt
./juicebox.sh hiccups -r 10000 -f 0.1 -p 1 -i 5 sporozoite.hic hiccups_loops.txt

I'm not sure if it's from the initial or final list--I'm only doing one resolution.

Kate

On Mon, Mar 14, 2016 at 7:06 PM, Suhas Rao <rao....@gmail.com> wrote:
Also, what list does this entry appear in? The final merged list (i.e. after post-processing) or the first list of enriched pixels?


On Monday, March 14, 2016 at 6:39:19 PM UTC-7, Muhammad Shamim wrote:
Hi Kate,

What were the exact parameters/flags for the HiCCUPS run?
On Mon, Mar 14, 2016 at 7:41 PM, Kate Cook <kate...@uw.edu> wrote:
Hi,

When I run HICCUPS with a loose FDR threshold of 0.9, I get this loop reported:

chr1    x1    x2    chr2    y1    y2    color    observed    expectedBL    expectedDonut    expectedH    expectedV    fdrBL    fdrDonut    fdrH    fdrV    numCollapsed    centroid1    centroid2    radius
chr5    430000    440000    chr5    600000    610000    0,255,255    35.0    12.019885    11.546831    12.39225    10.399725    0.001752831    0.0017599224    0.09893618    0.0017469053    2    435000    600000    25000

But when I run it with a threshold of 0.1, that loop is absent, even though all four FDR values are <0.1.

Am I missing something about how that FDR threshold is used? I thought that all four (doughnut, BL, H, V) had to be below the threshold.

Thanks,
Kate

--
You received this message because you are subscribed to the Google Groups "3D Genomics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics...@googlegroups.com.




--
Kate Cook
Postdoctoral Fellow, Noble Lab
Department of Genome Sciences
University of Washington
kate...@uw.edu

Suhas Rao

unread,
Mar 14, 2016, 11:04:50 PM3/14/16
to 3D Genomics, rao....@gmail.com
What is the name of the output file you're looking at?

On Monday, March 14, 2016 at 7:55:06 PM UTC-7, Kate Cook wrote:
These are the commands I'm using:

./juicebox.sh hiccups -r 10000 -f 0.9 -p 1 -i 5 sporozoite.hic hiccups_loops.txt
./juicebox.sh hiccups -r 10000 -f 0.1 -p 1 -i 5 sporozoite.hic hiccups_loops.txt

I'm not sure if it's from the initial or final list--I'm only doing one resolution.

Kate
On Mon, Mar 14, 2016 at 7:06 PM, Suhas Rao <rao....@gmail.com> wrote:
Also, what list does this entry appear in? The final merged list (i.e. after post-processing) or the first list of enriched pixels?


On Monday, March 14, 2016 at 6:39:19 PM UTC-7, Muhammad Shamim wrote:
Hi Kate,

What were the exact parameters/flags for the HiCCUPS run?
On Mon, Mar 14, 2016 at 7:41 PM, Kate Cook <kate...@uw.edu> wrote:
Hi,

When I run HICCUPS with a loose FDR threshold of 0.9, I get this loop reported:

chr1    x1    x2    chr2    y1    y2    color    observed    expectedBL    expectedDonut    expectedH    expectedV    fdrBL    fdrDonut    fdrH    fdrV    numCollapsed    centroid1    centroid2    radius
chr5    430000    440000    chr5    600000    610000    0,255,255    35.0    12.019885    11.546831    12.39225    10.399725    0.001752831    0.0017599224    0.09893618    0.0017469053    2    435000    600000    25000

But when I run it with a threshold of 0.1, that loop is absent, even though all four FDR values are <0.1.

Am I missing something about how that FDR threshold is used? I thought that all four (doughnut, BL, H, V) had to be below the threshold.

Thanks,
Kate

--
You received this message because you are subscribed to the Google Groups "3D Genomics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics+unsubscribe@googlegroups.com.

Kate Cook

unread,
Mar 15, 2016, 12:40:30 PM3/15/16
to Suhas Rao, 3D Genomics
hiccups_loops.txt contains these six loops when -f is 0.9:

chr1 x1 x2 chr2 y1 y2 color observed expectedBL expectedDonut expectedH expectedV fdrBL fdrDonut fdrH fdrV numCollapsed centroid1 centroid2 radius
chr12 1010000 1020000 chr12 1730000 1740000 0,255,255 7.0 1.9589589 1.8390803 1.515716 1.9977293 0.22487673 0.2167159 0.07090168 0.7047101 2 1015000 1740000 25000
chr12 1580000 1590000 chr12 2360000 2370000 0,255,255 6.0 2.4392655 2.372962 2.4750762 2.2167969 0.75935024 0.79570866 0.7749153 0.65926415 2 1585000 2360000 25000
chr12 1010000 1020000 chr12 2190000 2200000 0,255,255 6.0 1.896892 1.9186133 2.2835777 1.7118393 0.51347184 0.5655289 0.7749153 0.5631627 2 1015000 2190000 25000
chr12 1220000 1230000 chr12 2710000 2720000 0,255,255 5.0 1.4460441 1.5115591 1.4947412 1.4972422 0.4675494 0.42156097 0.510472 0.5012743 2 1225000 2710000 25000
chr12 970000 980000 chr12 1580000 1590000 0,255,255 5.0 1.8311383 2.1143847 2.2381763 2.38415 0.7254416 0.8802418 0.8179747 0.753393 2 985000 1585000 30000
chr5 430000 440000 chr5 600000 610000 0,255,255 35.0 12.019885 11.546831 12.39225 10.399725 0.001752831 0.0017599224 0.09893618 0.0017469053 2 435000 600000 25000

and is blank when -f is 0.1.

Kate

To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics...@googlegroups.com.



--
Kate Cook
Postdoctoral Fellow, Noble Lab
Department of Genome Sciences
University of Washington
kate...@uw.edu

--
You received this message because you are subscribed to a topic in the Google Groups "3D Genomics" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3d-genomics/tFF28oQImR4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 3d-genomics...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/3d-genomics/efbf83f8-b3e5-4f32-b946-24095b014201%40googlegroups.com.

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

Suhas Rao

unread,
Mar 15, 2016, 5:59:46 PM3/15/16
to 3D Genomics
Hey Kate,

Note that hiccups_loops.txt (the final output file) contains the loop calls after post-processing, i.e. after clustering of close-by enriched pixels, and after application of enrichment thresholds, q-value sum thresholds. You can see the post-processing details in section VI.a.5.iii-vii of our Extended Experimental Procedures of Rao and Huntley, et al, Cell 2014. 

This is why you do not see the particular loop you pointed out when you look at the final loop list using -f=0.1. You can adjust the various post-processing thresholds from the defaults via the -t and -d flags. 

Also, there are some changes to the output files of HiCCUPS that will likely get pushed soon that will make it so that the list of enriched pixels called by HiCCUPS (without any postprocessing) is also outputted. But note that these cannot be regarded as high-confidence loop calls (as explained in the EEP of Rao and Huntley, et al). 

Cheers,
Suhas 
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics+unsubscribe@googlegroups.com.



--
Kate Cook
Postdoctoral Fellow, Noble Lab
Department of Genome Sciences
University of Washington
kate...@uw.edu

--
You received this message because you are subscribed to a topic in the Google Groups "3D Genomics" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3d-genomics/tFF28oQImR4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 3d-genomics+unsubscribe@googlegroups.com.

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

Kate Cook

unread,
Mar 15, 2016, 6:04:34 PM3/15/16
to Suhas Rao, 3D Genomics
Ah, thanks for explaining. I would definitely be interested getting the initial list as well--will there be a post on this newsgroup when that is pushed out?

Kate

Suhas 
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics...@googlegroups.com.



--
Kate Cook
Postdoctoral Fellow, Noble Lab
Department of Genome Sciences
University of Washington
kate...@uw.edu

--
You received this message because you are subscribed to a topic in the Google Groups "3D Genomics" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3d-genomics/tFF28oQImR4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 3d-genomics...@googlegroups.com.

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



--
Kate Cook
Postdoctoral Fellow, Noble Lab
Department of Genome Sciences
University of Washington
kate...@uw.edu

--
You received this message because you are subscribed to a topic in the Google Groups "3D Genomics" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3d-genomics/tFF28oQImR4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 3d-genomics...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/3d-genomics/95598393-f1da-454e-8a05-365422e41932%40googlegroups.com.

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

Muhammad Shamim

unread,
Jun 29, 2016, 8:35:16 PM6/29/16
to 3D Genomics
Yes, my apologies for the delay
The latest jars are available here:


On Tuesday, March 15, 2016 at 5:04:34 PM UTC-5, Kate Cook wrote:
Ah, thanks for explaining. I would definitely be interested getting the initial list as well--will there be a post on this newsgroup when that is pushed out?

Kate
On Tue, Mar 15, 2016 at 2:59 PM, Suhas Rao wrote:
Hey Kate,

Note that hiccups_loops.txt (the final output file) contains the loop calls after post-processing, i.e. after clustering of close-by enriched pixels, and after application of enrichment thresholds, q-value sum thresholds. You can see the post-processing details in section VI.a.5.iii-vii of our Extended Experimental Procedures of Rao and Huntley, et al, Cell 2014. 

This is why you do not see the particular loop you pointed out when you look at the final loop list using -f=0.1. You can adjust the various post-processing thresholds from the defaults via the -t and -d flags. 

Also, there are some changes to the output files of HiCCUPS that will likely get pushed soon that will make it so that the list of enriched pixels called by HiCCUPS (without any postprocessing) is also outputted. But note that these cannot be regarded as high-confidence loop calls (as explained in the EEP of Rao and Huntley, et al). 

Cheers,
Suhas 


On Tuesday, March 15, 2016 at 9:40:30 AM UTC-7, Kate Cook wrote:
hiccups_loops.txt contains these six loops when -f is 0.9:

chr1 x1 x2 chr2 y1 y2 color observed expectedBL expectedDonut expectedH expectedV fdrBL fdrDonut fdrH fdrV numCollapsed centroid1 centroid2 radius
chr12 1010000 1020000 chr12 1730000 1740000 0,255,255 7.0 1.9589589 1.8390803 1.515716 1.9977293 0.22487673 0.2167159 0.07090168 0.7047101 2 1015000 1740000 25000
chr12 1580000 1590000 chr12 2360000 2370000 0,255,255 6.0 2.4392655 2.372962 2.4750762 2.2167969 0.75935024 0.79570866 0.7749153 0.65926415 2 1585000 2360000 25000
chr12 1010000 1020000 chr12 2190000 2200000 0,255,255 6.0 1.896892 1.9186133 2.2835777 1.7118393 0.51347184 0.5655289 0.7749153 0.5631627 2 1015000 2190000 25000
chr12 1220000 1230000 chr12 2710000 2720000 0,255,255 5.0 1.4460441 1.5115591 1.4947412 1.4972422 0.4675494 0.42156097 0.510472 0.5012743 2 1225000 2710000 25000
chr12 970000 980000 chr12 1580000 1590000 0,255,255 5.0 1.8311383 2.1143847 2.2381763 2.38415 0.7254416 0.8802418 0.8179747 0.753393 2 985000 1585000 30000
chr5 430000 440000 chr5 600000 610000 0,255,255 35.0 12.019885 11.546831 12.39225 10.399725 0.001752831 0.0017599224 0.09893618 0.0017469053 2 435000 600000 25000

and is blank when -f is 0.1.

Kate
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics+unsubscribe@googlegroups.com.



--
Kate Cook
Postdoctoral Fellow, Noble Lab
Department of Genome Sciences
University of Washington
kate...@uw.edu

--
You received this message because you are subscribed to a topic in the Google Groups "3D Genomics" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3d-genomics/tFF28oQImR4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 3d-genomics+unsubscribe@googlegroups.com.

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



--
Kate Cook
Postdoctoral Fellow, Noble Lab
Department of Genome Sciences
University of Washington
kate...@uw.edu

--
You received this message because you are subscribed to a topic in the Google Groups "3D Genomics" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3d-genomics/tFF28oQImR4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 3d-genomics+unsubscribe@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages