molikd / otb Goto Github PK

https://github.com/mermaid-js/mermaid works natively in github now, diagram from otb should be made and added to README

capture multiple bams as a list, loop through and run flagstat once per bam. This is sanity check for the bams being the right size, having correct number of reads, etc.

The following tools require buildout of flag features in the run.nf file, which need to be bubbled up into otb.sh, and the otb.sh --help option as well as the wiki documentation.

• #9
• #10
• #11
• #12
• #13
• #14
• #15
• #16
• #17
• #18

yahs juicer_pre for yahs outputs, and juicebox tools juicer pre for fragments list from hicstuff.

something about the nextflow call in the otb script is evaluating the string wildcards in the hifi and hic data inputs.

slurm_usda config needs right sizing for various processes

The input bam for yahs should be alignments of the hic yeads to the contigs, not filtered hifi as it is currently.

The ragtag patch step with the ec reads is hella slow, this is the one before HiC is run. Either need to speed it up or remove it.

Need to build out bwa mem container https://github.com/lh3/bwa for #33

Currently the trio mode of hifiam is using the HiC data, but it should be using maternal and paternal reads (https://hifiasm.readthedocs.io/en/latest/trio-assembly.html)
Need to set up optional parental reads channels. I think renaming the hic reads as such will also help clarify input and use cases.

this line in gfa2fasta needs to be updated to the genomes of any of the HiFiASM modes, namely in phasing mode hic.p_ctg is the genome

file '*.bp.p_ctg.gfa.fasta' optional true into fasta_unoriented_ch, fasta_genome_ch, fasta_busco_ch, no_polish_yahs_align_genome_ch, fasta_fai_yahs_genome_ch

Following pages needed to be updated to new structure:

reports
results

In order to publish otb we need to run it on something, in this case the two lined spittle bug genome.

target journal is g3

Yahs to be added to otb, ragtag will need to be run on yahs output to align rebuild hap1 and hap2 outputs.

Check "homozygous", and "heterozygous" options for HiFiASM, possibly not reading the HiFi reads?

see the run.nf.

params.hicreadf = "$baseDir/data/.R1.fastq.gz"
params.hicreadr = "$baseDir/data/.R2.fastq.gz"
params.matreadf = "$baseDir/data/.R1.fastq.gz"
params.matreadr = "$baseDir/data/.R2.fastq.gz"
params.patreadf = "$baseDir/data/.R1.fastq.gz"
params.patreadr = "$baseDir/data/.R2.fastq.gz"

otb needs correct arguments

Ragtag could potentially generate false positives and incorrectly capture inversions between haplotypes. Inverted haplotype of chromosome 4 on locust genome could be a good case study.

Update hifiadapterfilt to 2.0 https://github.com/sheinasim/HiFiAdapterFilt

is running currently on HiC data

otb.sh and run.nf conversion to use multiple HiFi reads files instead of just two

May be worthwhile to use nf-core templates when possible, such as https://github.com/nf-core/modules/tree/master/modules/busco

process K_mer_counting filt_fastq_ch needs to be flattened so that all input hifi reads are k-mer counted.

Thank you for your work. I'm trying to use otb(only the best), but I see that the input of --readbam in wiki is bam file. I wonder if this can be in fasta format. Most assembly software uses fasta format files, and I can't find my original bam file.

I tried to run it, but an error occurred. I don't know if I didn't understand it well. I hope you can give me some advice.

./otb.sh --runner none --mode homozygous --threads 40 -f 00.data/hic_R1.fastq -r 00.data/hic_R2.fastq --polish-type simple --bam hifi.fatsa

[Sun Jul 3 20:49:30 CST 2022]: checking runner
[Sun Jul 3 20:49:30 CST 2022]: none being used
[Sun Jul 3 20:49:30 CST 2022]: checking polishing
[Sun Jul 3 20:49:30 CST 2022]: simple/ragtag polishing
[Sun Jul 3 20:49:30 CST 2022]: building run parameters
[Sun Jul 3 20:49:30 CST 2022]: not running busco
[Sun Jul 3 20:49:30 CST 2022]: reads file(s) not given, exiting

thanks for your help！

if multiple fastq files, hifiasm is running once per fastq rather than using all fastqs for one hifiasm. Seems to just retain whatever is the last hifiasm to finish (overwriting other ones).

nf-core?

Soemthing from the hifasm git repo?

https://zenodo.org/ button needs to be added to otb before publishing.

(base) [scott.geib@ceres ~]$ pwd
/home/scott.geib

-rwxrwxr-x. 1 scott.geib scott.geib 438M Sep 8 20:07 dmolik-blobtools.img
-rwxrwxr-x. 1 scott.geib scott.geib 521M Sep 8 19:52 dmolik-genomescope2.img
-rwxrwxr-x. 1 scott.geib scott.geib 74M Sep 8 19:53 dmolik-hifiasm.img
-rwxrwxr-x. 1 scott.geib scott.geib 59M Sep 8 19:53 dmolik-jellyfish.img
-rwxrwxr-x. 1 scott.geib scott.geib 400M Sep 8 19:56 dmolik-pbadapterfilt.img
-rwxrwxr-x. 1 scott.geib scott.geib 392M Sep 8 19:58 dmolik-ragtag.img
-rwxrwxr-x. 1 scott.geib scott.geib 122M Sep 8 20:00 dmolik-shhquis.img

large datasets can currently cause out of memory errors while running hifiasm, limit (in part) with -f

when --haplotype-scaffoldings or something.

better to bubble up the "dash l" option from hifiasm to the user.

This philosophical question about procedures for major tool changes and versioning for otb genomes, how do we want to handle this?

https://github.com/vgl-hub/gfastats