parameters for Denovo assembly

[nareshmvr]

for bacterial genome

how the parameters bubble size , K-mer , Length size , contig length , mimatch cost , insertion cost , deletion cost , length fraction , similarity fraction effects for Illumina data in Denovo Assembly

[kcchan]

The documentation of the program should explain each of these parameters in great detail.

[diptarka]

Guys, I am new to NGS and have a problem. I have been provided with fastq sequences of an yeast which are 66 in numbered and of the order abcd_2_NoIndex_L002_R1_001.fastq upto abcd_2_NoIndex_L002_R33_001.fastq and similarly abcd_2_NoIndex_L002_R2_001.fastq upto abcd_2_NoIndex_L002_R2_033.fastq and have been told it's a paired end data. How do i carry out de novo assembly of the same as i am interested in finding out the presence of certain genes of interest. I had been told that velvet is one such assembler for these, but it accepts only one pair of r interleaved files for processing. How do i modify my files according to velvet standards. Kindly guide me on the commands and steps required for the same. Thanks a ton in advance.

[jgibbons1]

diptarka,
Within your velvet directory there should be a perl script called shuffleSequences_fasta.pl

This will order your 2 fastq files into a single, correctly ordered file.

Usage From Velvet Manual:
./shuffleSequences_fasta.pl forward_reads.fa reverse_reads.fa output.fa

[diptarka]

@jgibbons1:
actually i have in total 66 fastq files. i know the shufflesequences.pl script. what i am asking is the 33 fastq files for the forward end that i have, can i combine them all into 1? and so for the rest 33 for paired end?otherwise the shufflesequences.pl script won't run. I was asking whether the procedure is correct or not? cos, i just don't have two fastq files, i have 33+33 fastq files that are paired end.

[jgibbons1]

Ahhh...I see, sorry I miss understood.

It sounds like you have 2 options:

(1) Use the shufflesequences.pl script on each of the 33 pairs
Velvet can handle multiple input files, so you can list all files in velveth, or simply use a wildcard -shortPaired *.fastq

(2) Merge the 33 forward and 33 reverse fastq files and then run the shufflesequences.pl script

You can merge files in a unix environment using the 'cat' command. If you go with this route, be sure to keep the files in the same order.

[diptarka]

@jgibbons1:
i can merge the 33 files into one single abcd_1.fastq and similarly the rest into abcd_2.fastq. But, the read info won't be hampered is it?I have actually done that, but was unsure whether further processing via this method will hamper or not.

[kcchan]

Merging the files shouldn't change any information in the files themselves. The key part is making sure the files in both reads are merged in the same order to maintain their pairing.

[diptarka]

by same order you mean sequentially right from 1 to 33 for one end and 1 to 33 for the other. I guess the cat command in linux will merge sequentially.

[diptarka]

Do, any filtering of the reads necessary before the files are merged and interleaved for velvet to be run?any qc checks required?

[jgibbons1]

Exactly what kcchan said. Just make sure you merge the files in the same order (1, 2, 3...33) in both read sets.

Filtering is a different story. If you are trimming the file as a whole, (i.e. trim 10 bp from beginning and end) there should be no problem. If you are filtering reads individually you need to be careful that bad reads are removed from both files. A quick way to check if this worked correctly is by checking the number of reads in each file at the end of filtering:

wc -l yourfile.fastq | awk '{seq=$1/4} END {print seq}'