HGA tool version 1.0
This tool helps to apply the Hierarchical Genome Assembly (HGA) method. 
The tool will apply:
1. Partitioning a given reads dataset into a given number of partitions.
2. Assembling each partitions using a pre-specified assembler (Velvet or SPAdes in this version) and using a given kmer size.
3. Merging all the assemblies of the partition.
4. Combining all the assemblies of the partition (using velvet with kmer value of 31).
5. Finaly, re-assembling the whole dataset with the merged contigs or the combined contigs, using a given kmer size.

--Software version--
This is version 1.0.0, supporting complete HGA assembly using SPAdes and velvet to assemble the partitions, for now.

--Prerequisite--
Python v2.7.6. 
Velvet should be installed with option [make 'MAXKMERLENGTH=111' 'LONGSEQUENCES=1']. 'LONGSEQUENCES=1' to allow Velvet to accept contigs (long sequences) as input;
while 'MAXKMERLENGTH=111' to setup assembly using kmer of size up to 111 as the default max kmer length is 31.

--Installation--
The tool is built using python; there is no need for installation.

--Running HGA Tool--
Run the command 
python HGA.py [options]

where the options are:
--Parameters--
 -velvet  path      path to the directory where velvet bineries (velveth, velvetg) are located
 -spades  path      path to the directory where spades.py file is located
 -PA      string    select "SPAdes" or "velvet" to choose which assembler will be used to assemble the partitions
 -P12     file      file of the interlaced fastq file of paired reads that will be used in the partiton step, usually raw or clean (corrected) reads
 -R12     file      file of the interlaced fastq file of paired reads that will be used in the re-assembly step, usually raw or clean (corrected) reads
 -ins     int       insert size of the fragments
 -std     int       standard deviation of the inset size
 -P       int       number of partitions
 -Pkmer   int(odd)  kmer value for assembling the partitions
 -Rkmer   int(odd)  kmer value for re-assembly step
 -t       int       number of threads to be used for re-assembly step using SPAdes assembler, default 1.   
 -out     Path      Output path
 -h                 print help menue

-Output formats--
After HGA finish running, the following will be at the given output directory:
	- All, 1-p, partitioned fastq files; and their assemblies each in folder under the name of the partition.
	- ./combined_contigs/combined_contigs.fasta; fasta file of merging all contigs of the partitions assemblies.
	- ./merged_contigs/merged_contigs.fasta, fasta file of combining all contigs of the partitions assemblies.
	- ./HGA_merged/contigs.fasta (as well scaffolds.fasta); contigs of HGA(merged contigs) flow.
	- ./HGA_combined/contigs.fasta (as well scaffolds.fasta); contigs of HGA(combined contigs) flow.
	- ./HGA.log a log file of the run of the script.


--Samples tests--
We have created two scripts that run samples of HGA method. There are two scripts 
"sample_test_cholera.sh" and 
"sample_test_axonopodis.sh".


Each script runs as follows:
1. Download the dataset correspondent to the genome (V. cholera or X. axonopodis) from GAGE-B website.
2. Extract the dataset.
3. Download the script interleave_fastq.py which will be used to interleave to fastq paired reads; from
https://gist.githubusercontent.com/ngcrawford/2232505/raw/338758f7fcca8ad24340730b96ba645c46fa1b0e/interleave_fastq.py
4. Interleave the paired reads into one interlaced fastq file, since HGA takes as an input an interlaced fastq file.
5. Download and install locally Velvet assembler using the recommended installation command: 
make 'MAXKMERLENGTH=111' 'LONGSEQUENCES=1' 
6. Download and install locally SPAdes assembler.
7. Run HGA method using the fastq file and the local Velvet and SPAdes installations.
8. Delete all temporary files and leave only HGA result files and a log file that stores the output of all commands in the scripts.

For V. cholera HiSeq test:
The script will create a folder under the name "Cholera_test", which contains:
./run.log     this file contains the output of all commands used in the script.
./HGA 	      this folder contains the files of HGA which are:
	./HGA/HGA.log
	./HGA/HGA_combined/contigs.fasta
	./HGA/HGA_merged/contigs.fasta
	./HGA/part_{x}_assembly/contigs.fasta

To run the samples test, "HGA.py" must be in the same directory where the sample script is stored, then run:
bash sample_test_cholera.sh 
or 
bash sample_test_axonopodis.sh
 
The reported results:
HGA Settings: Number of partitions is 4 (each partition is ~25x of coverage), kmer used to assemble each partitions is 31 using Velvet assembler, and kmer used for re-assembly is 51.
Server specifications: Linux platform, AMD Opteron(tm) 2.4GH, 256GB Memory, 64 core. Python v2.7.6. 
Time: 100-120 minutes.
Thread: 1
N50:
./HGA/HGA_combined/contigs.fasta         246,465   
./HGA/HGA_merged/contigs.fasta           356,011       
./HGA/part_1_assembly/contigs.fasta       19,443


For X. axonopodis HiSeq test:
The script will create a folder under the name "Axonopodis_test", which contains:
./run.log     this file contains the output of all commands used in the script.
./HGA 	      this folder contains the files of HGA which are:
	./HGA/HGA.log
	./HGA/HGA_combined/contigs.fasta
	./HGA/HGA_merged/contigs.fasta
	./HGA/part_{x}_assembly/contigs.fasta
 
The reported results:
HGA Settings: Number of partitions is 16 (each partition is ~15x of coverage), kmer used to assemble each partitions is 31 using Velvet assembler, and kmer used for re-assembly is 51.
Server specifications: Linux platform, AMD Opteron(tm) 2.4GH, 256GB Memory, 64 core. Python v2.7.6. 
Time: 400-450 minutes.
Thread: 1
N50:
./HGA/HGA_combined/contigs.fasta         118,022   
./HGA/HGA_merged/contigs.fasta           115,197       
./HGA/part_1_assembly/contigs.fasta       12,932



--Contacts--
For more queries or questions, please contact
Anas Al-okaily, anas.al-okaily@uconn.edu


Last update: 12/2015