This repository is a stand-alone copy of RAmbler. Anyone can run RAmbler locally in their machines just by cloning this.
You will need conda installed in your machine to proceed. If you don't have conda installed, please follow conda installation. If you already had it installed, please make sure it is updated.
Create a conda environment first:
conda create -n rambler python=3.8 numpy pandas matplotlib seaborn xlsxwriter biopython
This will create a conda environment named rambler with the afforementioned packages. Then activate the environment with:
conda activate rambler
Finally, install the following packages inside the newly created environment using the commands below:
conda install -c bioconda jellyfish conda install -c bioconda minimap2 conda install -c bioconda bwa conda install -c bioconda samtools conda install -c bioconda pysam conda install -c bioconda hifiasm conda install -c bioconda seqkit
Currently, RAmbler only takes reads in .fasta format. If you have the reads in a different file format, please convert them to .fasta before executing the assembly pipeline.
mapReads.sh contains the required commands to perform step A (Determine the reads corresponding to repetitive regions)
extractUnikmers.sh contains the required commands to perform step B (Determine unikmers)
run.sh contains the required commands to execute RAmbler's steps C, D, E and F (the assembly pipeline)
stitch.sh contains the required commands to stitch the repeat resolved assembly back to the original draft/reference genome
If you already have the HiFi reads corresponding to the repetitive regions and the unikmers, directly execute run.sh to obtain the repeat resolved assembly.
If you don't have them, then first execute mapReads.sh by providing the current draft genome and the complete set of HiFi reads to it. Follow the detailed steps mentioned in mapReads.sh to extract reads mapping to the repetitive regions. For executing mapReads.sh, you will also need a local copy of NucFreq. Next, for extracting the unikmers, execute extractUnikmers.sh. Follow the detailed instructions mentioned in extractUnikmers.sh.
If you have already executed run.sh and obtained the repeat resolved assembly, execute stitch.sh to merge the newly reconstructed repeats with the draft genome.
Open the terminal and follow the steps below to run RAmbler from scratch:
cd ~ git clone https://github.com/sakshar/rambler.git git clone https://github.com/mrvollger/NucFreq.git
We have provided a test dataset inside data directory:
data/10.fasta contains the HiFi reads generated with a 10x coverage depth
data/10.kmers contains the unikmers
data/10000_5_100.fasta contains the true reference genome for the above HiFi reads and unikmers. It contains 5 copies of a 10 Kb repeat unit with a mutation rate of 1 per 100 bases along with 50 Kb upstream and downstream flanking regions
Always run RAmbler from inside the rambler/scripts directory. So, to run RAmbler directly when you have the required HiFi reads and unikmers, execute the following command:
cd ~ cd rambler/scripts bash run.sh -r read_file -u unikmer_file -o output_directory -s assembly_length [-k kmer_length] [-l tolerance] [-h threshold] required: -r path to the fasta file containing the HiFi reads extracted in step A -u path to the file containing the list of unikmers extracted in step B -o path to the directory where RAmbler will put all the generated outputs -s estimated length of the repeat resolved part of the genome in bp optional: -k length of kmers used (default: 21) -l tolerance parameter (default: 15) -h threshold parameter (default: 15)
The repeat resolved assembly will be generated in a file named rambler.fasta inside output_directory/assembly directory.
Next, execute the following command to merge rambler.fasta with the draft reference genome for obtaining the final assembly:
bash stitch.sh -g draft_genome_file -n contig_id -b start_position -e end_position -o output_directory -s contig_length required: -g path to the fasta file containing the draft/reference genome -n id of the contig that is being re-assembled -b approximate start position of the repeat region in the contig sequence from the draft genome (if stitching at the end, put -1) -e approximate end position of the repeat region in the contig sequence from the draft genome (if stitching at the beginning, put -1) -o path to the directory where RAmbler will put all the generated outputs (always put the same one as run.sh) -s estimated length of the repeat resolved contig in bp
bash run.sh -r ../data/10.fasta -u ../data/10.kmers -o ../output -s 150000
This will create an output directory rambler/output and the reconstruced repeats will be in rambler/output/assembly/rambler.fasta.
bash stitch.sh -g ../data/10000_5_100.fasta -n draft -b 50000 -e 100000 -o ../output -s 150000
This will create a sub-directory final inside the previously created output directory rambler/output and the repeat reconstructed contig will be in rambler/output/final/rambler_merged.fasta.
- Sakshar Chakravarty, Glennis Logsdon, Stefano Lonardi. RAmbler: de novo genome assembly of complex repetitive regions. bioRxiv 2023.05.26.542525; doi: https://doi.org/10.1101/2023.05.26.542525
or - Sakshar Chakravarty, Glennis Logsdon, and Stefano Lonardi. 2023. RAmbler: de novo genome assembly of complex repetitive regions. In 14th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics (BCB ’23), September 3–6, 2023, Houston, TX, USA. ACM, New York, NY, USA, 1 page. https://doi.org/10.1145/3584371.3612971
- How to calculate the lower and upper bounds of the window for extracting unikmers from the generated excel file?
While calculating for the mean and the standard deviation from the excel file, please ignore the first 5 rows (which are mostly k-mers due to sequencing errors) and also ignore the k-mers that appear way higher than the expected coverage depth (suppose, if the expected coverage depth is 100x, ignore rows after row 250). For example, with a read set having a coverage depth of 100x, try to calculate the mean and the standard deviation with values from rows 6 to 250. This is a trick required to discard k-mers coming from sequencing errors and the k-mers which appear way higher than expected that can throw the values of mean and standard deviation way off.
Note: the scripts, freq_distribution_kmers.py and extractUnikmers.sh, have been updated to fully automate step B for the extraction of unikmers
A stand-alone conda package for RAmbler with a single command execution feature