ft fire output result is empty

[bioluria]

Hello Mitchell! When I run FIRE version 0.1.2, I encountered some problems:

I download the test data GRCh38.m84039_241007_222736_s2.hifi_reads.bam from the following articles.
Vollger MR, Korlach J, Eldred K C, et al. Synchronized long-read genome, methylome, epigenome and transcriptome profiling resolve a Mendelian condition [ J ]. Nature Genetics, 2025 : 1-11.

config.tbl and config.yaml files are as follows:
$cat config.tbl
sample bam
Demo GRCh38.m84039_241007_222736_s2.hifi_reads.bam

$cat config.yaml
ref: GRCh38_GIABv3.fasta
ref_name: hg38
manifest: config.tbl
max_t: 12

$ pixi run fire-configfile config.yaml-cores 12
The output is empty when running to the following code
https://github.com/fiberseq/FIRE/blob/main/workflow/rules/apply-model.smk
In rule fire

samtools view -@ {threads} -u -F {params.flag} {input.bam} \
    | {FT_EXE} fire -F {params.flag} -t {threads} \
    {params.use_ont} \
    --min-msp {params.min_msp} \
    --min-ave-msp-size {params.min_ave_msp_size} \
    --skip-no-m6a \
    -  - 

After testing, I found that this step can be passed after ft fire uses the default parameters, that is, removing the -min-msp, -min-ave-msp-size and -skip-no-m6a options.

However, continue to run the process and find that in rule fire_sites_chrom

samtools view- @ { threads } -u { input.cram } { wildcards.chrom } \ 
    | { FT _ EXE } fire -t { threads } -extract - 

The result generated by is empty

My questions are as follows :

1. I don 't know whether it is necessary to preprocess the bam files after pbmm2 alignment before ft fire analysis to adapt to -min-msp, -min-ave-msp-size and-skip-no-m6a. That is, whether it is appropriate to remove these options, and whether there is a way to verify whether the bam entered into the FIRE process is valid.
2. It is not clear what is the role of ft fire in rule fire and rule fire_sites_chrom.
3. How to use GRCh38.m84039_241007_222736_s2.hifi_reads.bam to generate the final results

[mrvollger]

This is new data generated by PacBio using their new chemistry. The data in the paper you referenced can be found on SRA.

Nevertheless, after looking at this BAM, it looks like PacBio did not run nucleosome calling on it, which needs to be done before FIRE. See:
fiberseq/fibertools-rs#76

Cheers,
Mitchell