-
Notifications
You must be signed in to change notification settings - Fork 10
ChIP Benchmarking
The availability of ChIP data simulators was lacking, and/or too challenging to quickly use.
As a baseline benchmark, prior published results from ENCODE (human) and GEO (mouse) were used to compare against. Where possible parameters were matched; however, differences in mapping and calling are expected. The pipeline utilized by ngs-ops was derived from NF-core, which is a heavily tested and developed workflow. Here we aim to not fully reproduce published results, but to provide a sanity check that called peaks are within the realm of expectation.
A mouse study was selected from GEO, which had several samples each with its own input control. Data are SE, and narrowPeak files are generated with FDR q=0.01.
Intersect 'Truth' against 'Test'
singularity run intervene_check/intervene_0.6.5--pyh3252c3a_1.sif intervene venn --input \
CHIP_mouse_nxf_bench/immuno_precip_samples/adult_untreated_R1_vs_adult_untreated_input_R1/macs2/adult_untreated_R1_peaks.narrowPeak \
/projects/omics_share/mouse/GRCm38/supporting_files/benchmarking_data/CHIP/peak_files/GSM6592905_Adult_m_peaks.narrowPeak \
--names=ngs-ops_adult_untreated,geo_adult_untreated \
--colors=#4285F4,#80868B \
-o adult_untreated
singularity run intervene_check/intervene_0.6.5--pyh3252c3a_1.sif intervene venn --input \
CHIP_mouse_nxf_bench/immuno_precip_samples/aged_control_IgGTreated_R1_vs_aged_control_IgGTreated_input_R1/macs2/aged_control_IgGTreated_R1_peaks.narrowPeak \
/projects/omics_share/mouse/GRCm38/supporting_files/benchmarking_data/CHIP/peak_files/GSM6592906_Control_m_peaks.narrowPeak \
--names=ngs-ops_aged_untreated,geo_aged_untreated \
--colors=#4285F4,#80868B \
-o aged_untreated
singularity run intervene_check/intervene_0.6.5--pyh3252c3a_1.sif intervene venn --input \
CHIP_mouse_nxf_bench/immuno_precip_samples/aged_anti-IL17A_F_treated_R1_vs_aged_anti-IL17A_F_treated_input_R1/macs2/aged_anti-IL17A_F_treated_R1_peaks.narrowPeak \
/projects/omics_share/mouse/GRCm38/supporting_files/benchmarking_data/CHIP/peak_files/GSM6592907_Treated_m_peaks.narrowPeak \
--names=ngs-ops_aged_treated,geo_aged_treated \
--colors=#4285F4,#80868B \
-o aged_treated
Adult mouse control sample:
Aged mouse untreated:
Aged mouse treated:
A human study was selected from ENCODE, which had several samples with a shared replicated input control. Data are paired end, and narrowPeak file were generated with FDR q=0.1.
Intersect 'Truth' against 'Test'
singularity run intervene_check/intervene_0.6.5--pyh3252c3a_1.sif intervene venn --input \
CHIP_human_nxf_bench/immuno_precip_samples/H3K4me1_T0_R1_vs_H3K4me1_INPUT_R1/macs2/H3K4me1_T0_R1_peaks.narrowPeak \
/projects/compsci/omics_share/human/GRCh38/supporting_files/benchmarking_data/CHIP/peak_files/ENCFF477GAG.bed \
--names=ngs-ops_replicate_1,encode_replicate_1 \
--colors=#4285F4,#80868B \
-o human_replicate_1
singularity run intervene_check/intervene_0.6.5--pyh3252c3a_1.sif intervene venn --input \
CHIP_human_nxf_bench/immuno_precip_samples/H3K4me1_T0_R2_vs_H3K4me1_INPUT_R1/macs2/H3K4me1_T0_R2_peaks.narrowPeak \
/projects/compsci/omics_share/human/GRCh38/supporting_files/benchmarking_data/CHIP/peak_files/ENCFF082UGV.bed \
--names=ngs-ops_replicate_2,encode_replicate_2 \
--colors=#4285F4,#80868B \
-o human_replicate_2
Replicate 2:
Replicate 1: