raunaq-m / mlehaplo Goto Github PK

Hi,

Do I need to use error correction tools before I run the pipeline?

Thanks!

Hi!!
I am trying to run MLEHaplo in some synthetic data to compare some results between dbg-graph viral quasispecies reconstruction tools. Unfortunately, I am not being able to run MLEHaplo properlly because of the time.
I have attached here my script. The dataset has 600xcoverage which is 30 times smaller than the datasets used for the rest of the tools. However, MLEHaplo is stuck in the last part of the reconstruction. Maybe you can give me some hints about how to make it faster.
./multi-dsk/multi-dsk joint.fastq listOfKmer.txt
./multi-dsk/parse_results joint.solid_kmers_binary.55 > joint.60
perl construct_graph.pl joint.fastq joint.60 40 joint.55.graph "s"
perl construct_paired_without_bloom.pl -fasta joint.fastq -kmerfile joint.60 -thresh 40 -wr joint.60.pk.txt
perl dg_cover.pl -graph joint.55.graph -kmer joint.60 -paired joint.60.pk.txt -fact 15 -thresh 40 -IS 100 > joint.60.fact15.txt
perl process_dg.pl joint.60.fact15.txt > joint.60.fact15.fasta
perl get_paths_dgcover.pl -f joint.60.fact15.txt -w joint.60.fact15.paths.txt
screen -L -Logfile MLEHAPLO-log perl likelihood_singles_wrapper_parallel.pl -condgraph joint.55.cond.graph -compset joint.55.comp.txt -pathsfile joint.60.fact15.paths.txt -back -gl 10000 -slow > joint.60.smxlik.txt

Create a working example and a wrapper script so that the code can be run easily.

In the step

 perl construct_graph.pl "fastafile" "kmerfile" "threshold" "graphfile" "s"

One has to make a single FASTA file from paired end FASTQ. It appears that you have to run

perl fastqtofasta.pl <read file>

for each read, and then cat...is this OK (given that the reverse read is in the opposite direction)

I am asking because it seems to break the recursive git clone command.
While converting MLEHaplo into a local conda package (for convenience), conda build stumbled
over the folder. git clone --recursive url is the culprit.

git rm --cached MLEHaplo/

should fix it