Hi all! I looked back through the messages but can't seem to see the exact solution to my problem:

Error executing process > 'multiqc (1)'

Caused by:
  Process `multiqc` input file name collision -- There are multiple input files for each of the following file names: [long list of files here]

I have some quite large raw files, so I ended up splitting them into parts (with seqkit). So in my .input.tsv I have, say:

Sample_Name	Library_ID
NS0118	NS0118_part1
NS0118	NS0118_part2
NS0061_2025	NS0061_2025_part1
NS0061_2025	NS0061_2025_part2
NS0061_2023	NS0061_2023_part1
NS0061_2023	NS0061_2023_part2

Q1: Perhaps the duplicated sample IDs are causing problems for multiqc...? Also, in this case I have many sequencing runs of the same individual. Q2: what's the best way to merge the all data into one final genome?

Hi Pascale! can you please post the command you are using? Eager is built to handle multiple runs per library, and libraries per sample amd merge them together efficiently. It is exactly the answer for Q1 and Q2! So that is probably not the issue here. Are you using the latest version?

Hallo! After having a good nights sleep I realised the issue was that when creating the .tsv Excel had done the fun thing Excel does where it fills a series instead of copying cells down so I had sample names NS0061_2023, NS0061_2024, NS0061_2025, NS0061_2026, etc. Once I corrected this everything worked as planned in that I now have results for each sequencing run individually! Hooray! Is it possible to merge all the samples together (to create a consensus sequence/have more coverage for SNP calls, etc)? Or is this better done outside of the eager pipeline? To answer your questions: I'm not using the latest version, I'm on v`2.5.1` so that it will work with our cluster's NF version. If it's still needed, I've included the command below.

nextflow run nf-core/eager -r 2.5.1 -profile singularity -c 'nextflow.config' \
--input 'input_split_aoraki.tsv' \
--fasta 'Booted_eagle_rd2_consensus.fa' \
--bwa_index './' \
--fasta_index 'Booted_eagle_rd2_consensus.fa.fai' \
--seq_dict 'Booted_eagle_rd2_consensus.fa.dict' \
--outdir 'results/' \
-w 'working/' \
--complexity_filter_poly_g \
--preserve5p \
--mergedonly \
--dedupper 'dedup' \
--mapper 'bwamem' \
--run_bam_filtering \
--bam_mapping_quality_threshold 25 \
--bam_unmapped_type 'discard' \
--damage_calculation_tool 'damageprofiler' \
--damageprofiler_yaxis 0 \
--run_mtnucratio \
--mtnucratio_header '../Aquila_moorei_MK294166_mitochondrion.fasta' \
--max_memory 125GB \
--max_time 150h

I think to really answer your questions properly I would need to know the underlying structure and relationships between your data inputs. But if you put the same Sample_Id for multiple lines, they will be merged for consensus calls, genotyping etc. So essentially the “Sample_Id” is not equivalent to a “sample” in the wet-lab sense. It is the analysable unit consisting of all libraries merged together. Fr human popgen (where my background lies), I would use the individual ID as Sample_Id, library IDs as the library ID. Then deduplication and contamination estimates happen at the library level, before the data is merged for genotyping/consensus sequences etc