Chris:LabNotes/Moleculo/2013/2013-5-9: Difference between revisions

From ZhangLabWiki
Jump to navigation Jump to search
>Cjwei
>Cjwei
(Blanked the page)
Line 1: Line 1:
=Quick and Dirty Assembly Oncomelania Illumina contigs + moleculo data=
==Background==
*Dr. Zhang previously performed soapdenovo assembly on a smaller Illumina dataset, found in </media/LTS_15T/Kun_LTS/Snail/HB_species/Snail.HB.Soap.trimmed.K27.scafSeq>
*We wanted to quickly check how incorporating moleculo data will help improve assembly, so we used celera assembler to do OLC assembly on contigs>100bp and all of the moleculo data (without filtering, using default settings)
*Test run is found in my genemapster: </media/disk-2/cjwei/130419_Oncomelania_Illumina_Moleculo/Illumina_contigs+Moleculo_Assembly>


==Procedure==
*Convert Illumina contigs to fastq file format using converter utility (will create artificial fastq by assigning quality score of I to each base) -> this is required for inputting file into Celera
**Go into folder <~/software/Hybrid_error_correction_assembly_software/wgs_download_2>
**Use command:
java convertFastaAndQualToFastq ''input_file'' > ''output_file''
*Combined both Illumina contigs fastq file + Moleculo fastq file into new file: <Illumina_contig+Moleculo.fastq> (using cat command)
**It is necessary to have the correct ".fastq" file extension in order to continue forward
*Create .frg file for input to celera (this is the primary input for assembly that gives general data on the read file)
/media/2TB_store1/cjwei/software/Hybrid_error_correction_assembly_software/wgs_download_2/wgs-7.0/Linux-amd64/bin/fastqToCA -libraryname illumina_contigs_moleculo -reads Illumina_contig+Moleculo.fastq >Illumina_contig+Moleculo.frg
**Used "-reads" option because mimic single end reads for assembly
*RunCA to do assembly on combined dataset
~/software/Hybrid_error_correction_assembly_software/wgs-7.0/Linux-amd64/bin/runCA -d Illumina_contig+Moleculo_assembly -p Illumina_contig+Moleculo Illumina_contig+Moleculo.frg
**Without .spec file input, we are not doing correction as in normal hybrid assembly with pacbio+illumina data
==Results==
{| class="wikitable"
|-
! assembly !! n !! n:100 !! n:N50 !! min !! median !! mean !! N50 !! max !! sum
|-
| Illumina contig + Moleculo || 10322 || 10322 || 2812 || 1000 || 2458 || 3094 || 3682 || 17558 || 31.94e6
|-
| Illumina contig || 2167043 || 2167043 || 785219 || 100 || 132 || 152 || 148 || 4661 || 329.4e6
|}
*Results show a brief picture of the result of combining moleculo data with Illumina contigs.  Overall, there is a dramatic increase in N50 and max contig length.  However, there is a decrease in total bp assembled, probably because smaller contigs are being thrown out.
*Will need to run more detailed tests to better assemble the data.  Brief future plan:
**Optimize assembly of Illumina reads using complete dataset
**Run SEQuel on the assembled contigs to improve assembly accuracy
**Hybrid assembly of Illumina contigs + Moleculo data -> will need to optimize assembly parameters for celera or explore other assemblers
**Assembly analysis
*Will formulate more detailed plan and discuss with Dr. Zhang

Revision as of 20:34, 10 September 2013