Entering edit mode
6.4 years ago
mikysyc2016
▴
120
Hi, when i use macs2 diffpeak to find overlap and individual peaks of different TF, i find a weird things. I total have ~18000 peaks in condition1. But I get overlappeaks ~17000 in both condition(one bed file), and I get ~13000 peaks only in condition 1( another bed file). if 17000+13000 >> 18000? I do not understand how to explain this thing. Did someone can give me some suggestion?
macs2 bdgdiff --t1 '/home/yachen/Desktop/compare 129 and B6 ppara/B6_wt2/B6_wt2_treat_pileup.bdg' --c1 '/home/yachen/Desktop/compare 129 and B6 ppara/B6_wt2/B6_wt2_control_lambda.bdg' --t2 '/home/yachen/Desktop/compare 129 and B6 ppara/129a/129a_treat_pileup.bdg' --c2 '/home/yachen/Desktop/compare 129 and B6 ppara/129a/129a_control_lambda.bdg' --d1 27898981 --d2 36046837 -g 60 -l 120 --o-prefix diff_B6a_vs_129a
> track name="condition 2 (peaks)" description="unique regions in condition 2" visibility=1
chr1 3921530 3921680 diff_c1_vs_c2_cond2_1 1328844610706509347774191901104668672.00000
chr1 3924832 3925081 diff_c1_vs_c2_cond2_2 890016852594939748676134572142886912.00000
chr1 4026623 4026762 diff_c1_vs_c2_cond2_3 1376644783399102850724141617853235200.00000
chr1 4103233 4103445 diff_c1_vs_c2_cond2_4 902611436677182293366029879751999488.00000
chr1 5075121 5075272 diff_c1_vs_c2_cond2_5 1581434847060248510236660330297032704.00000
chr1 7415591 7415731 diff_c1_vs_c2_cond2_6 1506721422158697011656546888426979328.00000
chr1 8244113 8244304 diff_c1_vs_c2_cond2_7 1028782919306448365098449369013682176.00000
> track name="condition 1 (peaks)" description="unique regions in condition 1" visibility=1
chr1 3111442 3111586 diff_c1_vs_c2_cond1_1 1347560758765416526508252382036492288.00000
chr1 4484928 4485058 diff_c1_vs_c2_cond1_2 1635501095897932665382545844533723136.00000
chr1 4502979 4503142 diff_c1_vs_c2_cond1_3 1188531851806397262953375939235414016.00000
chr1 4509611 4509738 diff_c1_vs_c2_cond1_4 1693394857266680930677575467011145728.00000
chr1 4534261 4534425 diff_c1_vs_c2_cond1_5 1723906573388874298256900364812943360.00000
chr1 4797463 4797654 diff_c1_vs_c2_cond1_6 1017838420164890403767614801258217472.00000
chr1 5072948 5073080 diff_c1_vs_c2_cond1_7 2733623235386264950513463382178267136.00000
chr1 5823386 5823507 diff_c1_vs_c2_cond1_8 1581434847060248510236660330297032704.00000
chr1 7005263 7005383 diff_c1_vs_c2_cond1_9 1594613501156546211623295243908022272.00000
> track name="common (peaks)" description="common regions in both conditions" visibility=1
chr1 3890579 3890787 diff_c1_vs_c2_common_1 919969376346107438560724236743213056.00000
chr1 3893602 3894104 diff_c1_vs_c2_common_2 381182495874547323897889857084588032.00000
chr1 3921651 3921787 diff_c1_vs_c2_common_3 1693394857266680930677575467011145728.00000
chr1 4416962 4417274 diff_c1_vs_c2_common_4 613312891154684120952370293314158592.00000
chr1 4528206 4528333 diff_c1_vs_c2_common_5 1608013677195233796097841706404151296.00000
chr1 4540183 4540563 diff_c1_vs_c2_common_6 508919210082445201883807400629108736.00000
chr1 4661560 4661986 diff_c1_vs_c2_common_7 486904872846231713040065278588747776.00000
chr1 4819360 4819645 diff_c1_vs_c2_common_8 671416211013282615420334839540219904.00000
chr1 4847612 4847881 diff_c1_vs_c2_common_9 1039965340620036662235077349252005888.00000
I get 18437 peaks in con1 after macs2 peak calling and I get 31410 peaks in con2 after peak calling. but I get 8331 peaks in con1 after macs2 diffpeak, 4799 peaks in con2 after diffpeaks and 16263 pekas in common.bed. 8331+16263>18437 and 4799+16263< 31410? what is the reason? Thanks
Its not clear.
Whats the difference between 1 and 3. What is 2 ?
I add more content. Please see it.
Please share the command lines of how each file was produced, and please give some biological details on what these samples and conditions are. You cannot interpret an experiment without knowing the background story.
I add more content. Please see it.
I actually have neither a clue about the differential peak calling functionality of macs, nor if it is well-accepted in the community. If I was you, I would go for DiffBind or csaw (Bioconductor packages with excellent documentation) for differential analysis.
I do not have enough replicate, otherwise i will use diffbind.
...and you think that macs compensates for the lack of replicates or why do you use it? How many do you have?
I have rep_1 and rep_2. But I find when i want to find significantly differentially bound sites i need rep_1, rep_2 and rep_3. Do you have good suggestion? Thanks,