1. Input data
Upload control-group and experimental-group methylation files separately. Files may be uploaded as individual .txt files or as a .zip archive.
Each input file should be a whitespace-delimited table with five columns:
chromosome position methylated_read_count unmethylated_read_count context
Example:
chr1 10542 12 8 CpG chr1 10547 3 17 CHH
Replicate files should normally follow the naming pattern
1-wt.txt, 2-wt.txt, 1-mut.txt, 2-mut.txt, etc. The group names entered on the web page should match the suffixes used in the uploaded files.2. Basic parameters
| Web field | Core CLI argument | Meaning |
|---|---|---|
| Number of control/WT group replicates | --wt-reps | Number of biological replicates in the control / wild-type group. |
| Number of experimental/mutant group replicates | --mut-reps | Number of biological replicates in the experimental / mutant group. |
| Organism type | --biotype | 0: animal; 1: plant; 2: no p-value prefiltering for all contexts. |
| Control/WT group directory | --dir-wt | Directory containing the control / wild-type replicate files. File names should match the group name, for example 1-CK.txt. |
| Experimental/mutant group directory | --dir-mut | Directory containing the experimental / mutant replicate files. File names should match the group name, for example 1-D.txt and 2-D.txt. |
3. Threshold decision modes and options
MultiDMPcaller supports two threshold decision modes. In the recommended adaptive mode, q-value thresholds and final voting thresholds are determined automatically. In the user-defined mode, the manually specified thresholds are used instead.
| User-defined threshold | Adaptive counterpart | Default and meaning |
|---|---|---|
--q-cpg, --q-chg, --q-chh |
--auto-qvalue-twostep |
Manual context-specific DMP q-value thresholds. The plant-oriented default
values are 0.05 for CpG, 0.035 for CHG, and
0.045 for CHH. Users analyzing animal or other methylomes should
adjust these values according to their study design.
In the recommended adaptive mode, --auto-qvalue-twostep is enabled and
these manual q-value thresholds are kept for reference but are not used for final calling.
|
--vote-threshold |
--auto-dmp-vote-threshold, --auto-dmr-vote-threshold |
Manual final voting proportion across replicate-pair comparisons. The default manual
value is 0.6667, equivalent to 2/3. Decimal values and fractions
are accepted, for example 0.6667 or 2/3. In the recommended
adaptive mode, automatic DMP and DMR vote-threshold estimation is enabled and takes
priority over this manual value.
|
4. Other options
| Option | Recommended web default | Meaning |
|---|---|---|
--methy-diff-dmp |
0.0 |
Minimum absolute site-level methylation difference required for DMP support. Values are
in the 0-1 range, so 0.2 means 20%. This threshold is applied at the DMP
pair-support layer before final voting.
|
--methy-diff-dmr |
0.0 |
Minimum absolute regional methylation difference required for each pairwise DMR support. The regional difference is calculated from aggregated methylated and unmethylated reads within the candidate region. Values are in the 0-1 range. |
--dmr-q |
0.05 |
DMR q-value threshold used for final DMR calling. |
--dmp-lowdiff-strict-vote |
Enabled | Post-processes provisional final DMPs. Low-difference DMP candidates must satisfy a stricter support-count requirement, which helps reduce weakly supported low-difference calls. |
--dmp-lowdiff-cutoff |
0.3 |
Difference cutoff used by the low-difference strict-vote module. DMP candidates below this methylation-difference level are treated as low-difference candidates and are checked with a stricter voting rule. |
--processes |
4 on this web server |
Number of parallel worker processes. On the web server, each submitted job is capped at a maximum of 4 processes by default to avoid oversubscribing shared CPU resources. |
--dmr-engine |
cpp |
DMR candidate-generation backend. The cpp mode is recommended because it is
faster for large datasets. The python mode is kept as a compatible fallback.
|
--skip-dmr |
Disabled | Skip DMR analysis. The recommended default is disabled, meaning DMR detection is run. |
--skip-window |
Disabled | Skip sliding-window plotting and visualization. The recommended default is disabled, meaning visualization outputs are generated when possible. |
Recommended web configuration: adaptive q-value thresholding, adaptive DMP/DMR vote
thresholding, low-difference strict vote enabled,
--methy-diff-dmp 0.0, --methy-diff-dmr 0.0,
--dmr-engine cpp, and no skipping of DMR analysis or plotting.
5. Result files
After a job finishes, download the result archive. Important files are usually found under and_output/, including final DMP files such as CpG-final_significant_sites_DMPs.txt and final DMR files such as CpG-final_significant_regions_DMRs.txt.