hicFindTADs difficult to distinguish different TADs

[wangguiqian]

I haven't had a problem that hasn't been solved yet. Why is it so difficult to distinguish different TADs in the picture I took? The binsize about is 1K.
when I used the original V991.hic_matrix.h5 to obtain TADs, I also found that the boundaries and distinctions of TADs were difficult to distinguish. Is it because the binsize I initially set is wrong? Or is there an error somewhere? How do I know how to set the binsize size when running hicBuildMatrix to build an h5 matrix?

I use this command to display TADs:
hicFindTADs -m V991.hic_matrix.h5 --outPrefix TADs/V991_min9000_max_30000_step4500_fdr --minDepth 9000 --maxDepth 30000 --step 4500 --thresholdComparisons 0.05 --delta 0.01 --correctForMultipleTesting

hicPlotTADs --tracks Trach.ini --region Hic1:1-4000000 --dpi 300 --outFileName V991_Hic1_TADs.pdf --title "TADs_Hic1"

The V991_Hic1_TADs.pd show:

[joachimwolff]

Hi Kevin, Indeed the matrix does not look good. A 1k resolution requires a deep sequencing and even then, I would recommend to detect TADs with a resolution of 10k, maybe 25k. Hope this helps, Joachim kevin Wang ***@***.***> schrieb am Mi. 20. März 2024 um 12:03:

…

I haven't had a problem that hasn't been solved yet. Why is it so difficult to distinguish different TADs in the picture I took? The binsize about is 1K. when I used the original V991.hic_matrix.h5 to obtain TADs, I also found that the boundaries and distinctions of TADs were difficult to distinguish. Is it because the binsize I initially set is wrong? Or is there an error somewhere? How do I know how to set the binsize size when running hicBuildMatrix to build an h5 matrix? I use this command to display TADs: *hicFindTADs -m V991.hic_matrix.h5 --outPrefix TADs/V991_min9000_max_30000_step4500_fdr --minDepth 9000 --maxDepth 30000 --step 4500 --thresholdComparisons 0.05 --delta 0.01 --correctForMultipleTesting* *hicPlotTADs --tracks Trach.ini --region Hic1:1-4000000 --dpi 300 --outFileName V991_Hic1_TADs.pdf --title "TADs_Hic1"* The V991_Hic1_TADs.pd show: image.png (view on web) <https://github.com/deeptools/HiCExplorer/assets/48958211/61b1c139-2ca0-4230-82c9-6aadee49cb83> — Reply to this email directly, view it on GitHub <#896>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADGQCAF5RKCD7NTHHLOWKATYZFUILAVCNFSM6AAAAABE7HYXNSVHI2DSMVQWIX3LMV43ASLTON2WKOZSGE4TOMRVGQ4TQOI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[wangguiqian]

Hi Joachim,

When I use this command to merge binsize:
hicMergeMatrixBins -m 1C32.hic_matrix.h5 -o 1C32_hic_5kmatrix.h5 -nb 5

There is no problem when visualizing a chromosome, but it is still difficult to distinguish TAD when calling TAD.
hicPlotMatrix -m 1C32_hic_5kmatrix.h5 -o 1C32_5k-HiC_scaffold_1-plot.pdf --chromosomeOrder HiC_scaffold_1 --log1p --clearMaskedBins
The 1C32_5k-HiC_scaffold_1-plot.pdf show:

At present, the problem is that I don’t know how to set the binsize in terms of TAD and TAD boundary division. Or is there something wrong with the hicFindTADs command parameter setting? Is it a problem with the parameter settings of minDepth and maxDepth?

hicFindTADs -m 1C32_hic_5Kmatrix.h5 --outPrefix TADs_5K/1C32_min20000_max30000_step7500 --minDepth 20000 --maxDepth 30000 --step 7500 --thresholdComparisons 0.05 --delta 0.01 --correctForMultipleTesting fdr -p 10

The phone show:

Thank you very much for your help.
Best,
Kevin

[joachimwolff]

Hi Kevin, Reduce your resolution as I said to 10 kb or better 25kb. Call on 25kb and 10kb resolution TADs, do NOT use a 1kb or a 5kb. What you see in your 5kb plot that there are more or less no TADs visible, also applies for the TAD detection algorithm. It cannot detect on 1kb or 5kb resolution a TAD when your Hi-C data looks as you presented. That’s basically noise you have there. Best, Joachim

…

On 20. Mar 2024, at 14:40, kevin Wang ***@***.***> wrote: Hi Joachim, When I use this command to merge binsize: hicMergeMatrixBins -m 1C32.hic_matrix.h5 -o 1C32_hic_5kmatrix.h5 -nb 5 There is no problem when visualizing a chromosome, but it is still difficult to distinguish TAD when calling TAD. hicPlotMatrix -m 1C32_hic_5kmatrix.h5 -o 1C32_5k-HiC_scaffold_1-plot.pdf --chromosomeOrder HiC_scaffold_1 --log1p --clearMaskedBins The 1C32_5k-HiC_scaffold_1-plot.pdf show: image.png (view on web) <https://github.com/deeptools/HiCExplorer/assets/48958211/8b830d70-0971-451e-a38b-0d8ba70ceb11> At present, the problem is that I don’t know how to set the binsize in terms of TAD and TAD boundary division. Or is there something wrong with the hicFindTADs command parameter setting? Is it a problem with the parameter settings of minDepth and maxDepth? hicFindTADs -m 1C32_hic_5Kmatrix.h5 --outPrefix TADs_5K/1C32_min20000_max30000_step7500 --minDepth 20000 --maxDepth 30000 --step 7500 --thresholdComparisons 0.05 --delta 0.01 --correctForMultipleTesting fdr -p 10 The phone show: image.png (view on web) <https://github.com/deeptools/HiCExplorer/assets/48958211/9338a72e-e3e9-402b-b5a9-db819fa4b364> Thank you very much for your help. Best, Kevin — Reply to this email directly, view it on GitHub <#896 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADGQCAE3ZJ2BOWX6DXWF5DTYZGGTHAVCNFSM6AAAAABE7HYXNSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMBZGU4TCNBQGU>. You are receiving this because you commented.