search

aryeelab / hichipper

A preprocessing and QC pipeline for HiChIP data
MIT License
33 stars 12 forks source link

.rds not generated #55

Open aperreault opened 6 years ago

aperreault commented 6 years ago

Caleb,

I've gotten hichipper to analyze my HiChIP data if I include a ChIP-seq peak bed file in the .yaml (as opposed to finding peaks in the HiChIP data using MACS2 which was giving me lots of errors). However, I am still getting some errors. Below is an example of the terminal output: $ '/media/bryan/2Tb_Workspace/AP/HiChIP/hichipper_YY1.sh' Thu Oct 11 12:44:05 CDT 2018: Starting hichipper pipeline v0.7.3 Thu Oct 11 12:44:05 CDT 2018: Parsing user parameters Thu Oct 11 12:44:05 CDT 2018: .yaml file detected Anchors removed due to excessive size (likely at ends of chromosomes): 11594 Thu Oct 11 12:49:59 CDT 2018: Processing Rep1 Thu Oct 11 12:49:59 CDT 2018: Total_PETs=51917177 Thu Oct 11 12:50:01 CDT 2018: Mapped_unique_quality_pairs=37112928 Thu Oct 11 12:50:01 CDT 2018: Mapped_unique_quality_valid_pairs=1140248 Thu Oct 11 12:50:01 CDT 2018: Intersecting PETs with anchors Thu Oct 11 12:50:01 CDT 2018: Finished the anchor merging. Thu Oct 11 12:52:58 CDT 2018: Intrachromosomal_valid_small=652231 Thu Oct 11 12:52:59 CDT 2018: Intrachromosomal_valid_med=108725 Thu Oct 11 12:52:59 CDT 2018: Intrachromosomal_valid_large=66814 Thu Oct 11 12:52:59 CDT 2018: Total number of anchors used: 17698 Thu Oct 11 12:52:59 CDT 2018: Total number of reads in anchors: 5125794 Thu Oct 11 12:53:05 CDT 2018: Mapped_unique_intra_quality_anchor=54072 Thu Oct 11 12:53:05 CDT 2018: Mapped_unique_intra_quality_anchor_small=46117 Thu Oct 11 12:53:05 CDT 2018: Mapped_unique_intra_quality_anchor_med=6677 Thu Oct 11 12:53:05 CDT 2018: Mapped_unique_intra_quality_anchor_large=1278 Thu Oct 11 12:53:05 CDT 2018: Loop_PETs=6677 Thu Oct 11 12:53:05 CDT 2018: ['/usr/bin/Rscript', u'output1/qcReport.R', '/home/bryan/anaconda2/lib/python2.7/site-packages/hichipper', 'output1', '/media/bryan/2Tb_Workspace/AP/HiChIP/hichipper/YY1-hichipper', '0.7.3', 'Rep1'] /media/bryan/2Tb_Workspace/AP/HiChIP/hichipper/YY1-hichipperUsing count as value column: use value.var to override. Processing: Rep1 Error in { : task 1 failed - "missing value where TRUE/FALSE needed" Calls: %do% -> In addition: Warning message: In pbinom(q = v[1] - 1, size = v[2], prob = v[3]) : NaNs produced Execution halted

What does this mean? I've also noticed that the .rds file isn't generated. Do I need to call the diffloop_work.R script separately?

Thanks for all your help! -Andrea

caleblareau commented 6 years ago

So this is an error in the mango FDR calculation. It looks like you only have 6677 PETs in anchors between 5kb and 2Mb, which would suggest that this isn't a great dataset. The error then is happening in trying to fit the bias regression to assign a per-loop p-value.

aperreault commented 6 years ago

Okay, thanks! So if there were more Loop_PETS, this error probably wouldn't happen?