04_MAPPING

Step 0. Prep the data for mapping to be followed by binning

Yet another loop. Round and round we go... Within the 03_ASSEMBLIES directory:

# Reformat names

for SET in `cat set.txt`
do
    anvi-script-reformat-fasta $SET/$SET-contigs.fa -l 2500 --simplify-names -o $SET.fa
    anvi-gen-contigs-database -f 04_CONTIGS/$SET.fa -o 04_CONTIGS/$SET-CONTIGS.db
done

# HMM profiling
for SET in `cat set.txt`; do anvi-run-hmms --num-threads 20 -c 04_CONTIGS/$SET-CONTIGS.db ; done

# Run against NCBI COG database
for file in ./*.db; do anvi-run-ncbi-cogs --num-threads 16 -c $file ; done

Step 1. Index

# Build bowtie2 DB for each co-assembly
for SET in `cat set.txt`; do bowtie2-build 03_CONTIGS/$SET.fa 04_MAPPING/$SET --num-threads 20; done

Step 2. Actually do the mapping

while read line;      
    do         
    SET=$(echo $line | cut -d" " -f1);          
    samples=$(echo $line | cut -d" " -f2);          
    delimiter=",";          
    declare -a Smparray=($(echo $samples | tr "$delimiter" " "));         
    for samp in "${Smparray[@]}"; 
        do      
            bowtie2 --threads 40 \
            -x 05_MAPPING/$SET \
            -1 02_HUMAN_Decontam/$samp-cleanR1.fq \
            -2 02_HUMAN_Decontam/$samp-cleanR2.fq \
            --no-unal \
            -S 05_MAPPING/$samp.sam;     
            samtools view -F 4 -bS 05_MAPPING/$samp.sam > 05_MAPPING/$samp-RAW.bam;     
            samtools sort 05_MAPPING/$samp-RAW.bam -o 05_MAPPING/$samp.bam;     
            samtools index 05_MAPPING/$samp.bam;     
            rm 05_MAPPING/$samp.sam 05_MAPPING/$samp-RAW.bam;      
        done;     
    done < samples_in_sets.txt

Indexing seems... broken or something when I see the result on kjempefuru. I'm not sure why. Maybe this is something that SAGA is better suited for, this comapping. It's going really really slowly on kjempefuru. Perhaps doing the mappings as separate slurm jobs is the solution.

# copy contig data from kjempefuru to SAGA (within kjempefuru)

tocopy=$(find -type f -name '*.prefixed.fa') # everything's in subdirectories. Annoying.
scp $tocopy ede041@saga.sigma2.no:/cluster/projects/nn9836k/Metagenomics_AMOR_2020/04_CONTIGS/