pipeline does not start with requested folder structure

[BCArg]

Eventually what I want to do is to build de novo assemblies using flye and two rounds of racon and medaka each, to compare the results with a pipeline I build myself. Fastq's were generated with the high accuracy basecaller (therefore I am using flyehq), guppy version is 5.1.13+b292f4d

I have created the following folder structure, following the instructions on the README file:

(ont-assembly-snake) [ngs@vngs20x /nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake]$ tree
.
├── assemblies
│   ├── BORD1705_flyehq+racon2+medaka2
│   └── LEG829_flyehq+racon2+medaka2
└── fastq-ont
    ├── BORD1705.fastq
    └── LEG829.fastq

Then if I call the snakemake, also following the instructions on the README file:

(ont-assembly-snake) [ngs@vngs20x /nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake]$ snakemake -np -s /home/ngs/installed/snakemake/ont-assembly-snake/Snakefile --use-conda --cores 20

I get:

filtlong min. read length = 1000
Building DAG of jobs...
MissingInputException in line 64 of /home/ngs/installed/snakemake/ont-assembly-snake/Snakefile:
Missing input files for rule all:
assemblies/BORD1705_flyehq+racon2+medaka2/output.fa
assemblies/LEG829_flyehq+racon2+medaka2/output.fa
assemblies/BORD1705_flyehq+racon2+medaka2.fa
assemblies/LEG829_flyehq+racon2+medaka2.fa

[pmenzel]

Hi,

there is no keyword for running medaka multiple times, so you can only do +medaka.

From what I gathered, people usually run medaka only once.

You could actually try +medaka+medaka, for running it twice, that might work.

[BCArg]

Hi Peter,

thanks for the quick reply.

Still getting the error:

(ont-assembly-snake) [ngs@vngs20x /nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake]$ snakemake -np -s /home/ngs/installed/snakemake/ont-assembly-snake/Snakefile --use-conda --cores 20
filtlong min. read length = 1000
Building DAG of jobs...
MissingInputException in line 64 of /home/ngs/installed/snakemake/ont-assembly-snake/Snakefile:
Missing input files for rule all:
assemblies/.snakemake.fa
assemblies/.snakemake/output.fa

with the following folder structure (removed the '2' after medaka - which yes, according to my own pipeline, it does not change much the quality of the assembly):

.
├── assemblies
│   ├── BORD1705_flyehq+racon2+medaka
│   └── LEG829_flyehq+racon2+medaka
└── fastq-ont
    ├── BORD1705.fastq
    └── LEG829.fastq

[pmenzel]

Missing input files for rule all:
assemblies/.snakemake.fa
assemblies/.snakemake/output.fa

You probably accidently run snakemake inside the assemblies/ folder, which created the folder .snakemake in there.
Just remove that.

with the following folder structure (removed the '2' after medaka - which yes, according to my own pipeline, it does not change much the quality of the assembly):

I am curious how do you check the quality of your assemblies? Using any of the tools in https://github.com/pmenzel/score-assemblies or some others?

[BCArg]

I happened to initially have called the snakemake from another directory, in which case, it does not crash, but is shows only the rules for no input at all.

In my command above you can see that I called snakemake from /nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake and if I do a tree on that directory I get:

[ngs@vngs20x ~/installed/snakemake/ont-assembly-snake]$ tree /nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake
/nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake
├── assemblies
│   ├── BORD1705_flyehq+racon2+medaka
│   └── LEG829_flyehq+racon2+medaka
└── fastq-ont
    ├── BORD1705.fastq
    └── LEG829.fastq

4 directories, 2 files

As for the quality of the assemblies I use QUAST, great tool

[pmenzel]

try tree -a

[BCArg]

[ngs@vngs20x ~]$ tree -a /nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake/
/nexus/Analysis/microbiology/denovoAssemblies_ONT/ont-assembly-snake/
├── assemblies
│   ├── BORD1705_flyehq+racon2+medaka
│   ├── LEG829_flyehq+racon2+medaka
│   └── .snakemake
│       ├── auxiliary
│       ├── conda
│       ├── conda-archive
│       ├── incomplete
│       ├── locks
│       ├── log
│       ├── metadata
│       ├── shadow
│       └── singularity
├── fastq-ont
│   ├── BORD1705.fastq
│   └── LEG829.fastq
└── .snakemake
    ├── auxiliary
    ├── conda
    ├── conda-archive
    ├── incomplete
    ├── locks
    ├── log
    │   ├── 2022-06-14T142849.649163.snakemake.log
    │   ├── 2022-06-14T143237.621696.snakemake.log
    │   ├── 2022-06-14T143518.824018.snakemake.log
    │   ├── 2022-06-14T143732.727613.snakemake.log
    │   ├── 2022-06-15T102716.725747.snakemake.log
    │   ├── 2022-06-15T102838.335849.snakemake.log
    │   └── 2022-06-15T104224.438601.snakemake.log
    ├── metadata
    ├── shadow
    └── singularity

[pmenzel]

See, the folder assemblies/.snakemake is the one causing the trouble.

[BCArg]

Yes, I think that folder was created by snakemake when I called it from the assemblies directory.

Anyway, the hidden folder was empty and the pipeline starts after I removed it

Thanks for the support