Tool:New ultraperformant version of SICER/epic: epic2

Question

[endrebak852]

New version of the popular ChIP-Seq caller SICER out.

See https://github.com/biocore-ntnu/epic2 for more.

Performance:

CLI:

usage: epic2 [-h] [--treatment TREATMENT [TREATMENT ...]]
             [--control CONTROL [CONTROL ...]] [--genome GENOME]
             [--keep-duplicates] [--bin-size BIN_SIZE]
             [--gaps-allowed GAPS_ALLOWED] [--fragment-size FRAGMENT_SIZE]
             [--false-discovery-rate-cutoff FALSE_DISCOVERY_RATE_CUTOFF]
             [--effective-genome-fraction EFFECTIVE_GENOME_FRACTION]
             [--chromsizes CHROMSIZES] [--e-value E_VALUE] [--quiet]
             [--example]

epic2. (Visit github.com/endrebak/epic2 for examples and help.)

optional arguments:
  -h, --help            show this help message and exit
  --treatment TREATMENT [TREATMENT ...], -t TREATMENT [TREATMENT ...]
                        Treatment (pull-down) file(s) in one of these formats:
                        bed, bedpe, bed.gz, bedpe.gz or (single-end) bam, sam.
                        Mixing file formats is allowed.
  --control CONTROL [CONTROL ...], -c CONTROL [CONTROL ...]
                        Control (input) file(s) in one of these formats: bed,
                        bedpe, bed.gz, bedpe.gz or (single-end) bam, sam.
                        Mixing file formats is allowed.
  --genome GENOME, -gn GENOME
                        Which genome to analyze. Default: hg19. If
                        --chromsizes and --egf flag is given, --genome is not
                        required.
  --keep-duplicates, -kd
                        Keep reads mapping to the same position on the same
                        strand within a library. Default: False.
  --bin-size BIN_SIZE, -bin BIN_SIZE
                        Size of the windows to scan the genome. BIN-SIZE is
                        the smallest possible island. Default 200.
  --gaps-allowed GAPS_ALLOWED, -g GAPS_ALLOWED
                        This number is multiplied by the window size to
                        determine the number of gaps (ineligible windows)
                        allowed between two eligible windows. Must be an
                        integer. Default: 3.
  --fragment-size FRAGMENT_SIZE, -fs FRAGMENT_SIZE
                        (Single end reads only) Size of the sequenced
                        fragment. Each read is extended half the fragment size
                        from the 5' end. Default 150 (i.e. extend by 75).
  --false-discovery-rate-cutoff FALSE_DISCOVERY_RATE_CUTOFF, -fdr FALSE_DISCOVERY_RATE_CUTOFF
                        Remove all islands with an FDR below cutoff. Default
                        0.05.
  --effective-genome-fraction EFFECTIVE_GENOME_FRACTION, -egf EFFECTIVE_GENOME_FRACTION
                        Use a different effective genome fraction than the one
                        included in epic2. The default value depends on the
                        genome and readlength, but is a number between 0 and
                        1.
  --chromsizes CHROMSIZES, -cs CHROMSIZES
                        Set the chromosome lengths yourself in a file with two
                        columns: chromosome names and sizes. Useful to analyze
                        custom genomes, assemblies or simulated data. Only
                        chromosomes included in the file will be analyzed.
  --e-value E_VALUE, -e E_VALUE
                        The E-value controls the genome-wide error rate of
                        identified islands under the random background
                        assumption. Should be used when not using a control
                        library. Default: 1000.
  --quiet, -q           Do not write output messages to stderr.
  --example, -ex        Show the paths of the example data and print example command.

[Coralie]

Hi endrebak

I hope my question will not disturb on this topic. I would like to use epic2 on my chip seq dataset but i have an error. I have the same when i run the example :

epic2 -t /usr/local/lib/python3.7/dist-packages/epic2/examples/test.bed.gz -c /usr/local/lib/python3.7/dist-packages/epic2/examples/control.bed.gz > deleteme.txt Found a median readlength of 25.0

Using genome hg19.

Using an effective genome length of ~2510 * 1e6

Parsing ChIP file(s): /usr/local/lib/python3.7/dist-packages/epic2/examples/test.bed.gz terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc Abandon

Do you have any idea what's wrong?

Thanks

[ATpoint]

I suggest you open an issue at the GitHub repository of this tool: https://github.com/biocore-ntnu/epic2

[endrebak]

I would love to try to help you. Can you post an issue on the repo?

[koleilat.karim]

Hello,

Thank you for developing this tool.

I have CUT&RUN data I would like to call peaks on with epic2.

More specifically, I have + / - treatment samples with Ecoli DNA spiked in and would like a way to normalize/scale to the % ecoli read ratios between samples and Igg and between treatment vs vehicle.

is there a way to feed epic2 the scaled bedgraph files or the ratios for peak calling?

similar to the custom tutorial provided by macs2 using its subcommands?

https://github.com/macs3-project/MACS/wiki/Advanced:-Call-peaks-using-MACS2-subcommands

Thanks,

Karim

[GenoMax]

Please create a new post for this question. That author of this tools has not visited Biostars in 5+ years and will likely not see this.

Answer 1

[endrebak]

epic2 has been accepted into bioinformatics. I've also implemented the SICER-df and SICER-rb-df algorithms for differential enrichment. Will update with citation info later :)

[ben.kunfang]

Hi, endrebak

I was wondering if epic2 is specific for dispersed histone mark or there is any way to call sharp mark peak with epic2?

Thanks! Kun

[endrebak]

Macs2 is better for sharp peaks :) I've used epic2 for medium marks like PolII and H3K4me3, but not TFs with very sharp peaks.