Velvet Interpretation

[nxtgenkid10]

Hi I m trying to run velevt with the a bacterial genome with the following condition

k-mer=31 -cov_cutoff auto -exp_cov auto -ins_length 1000 -exp_cov 5 n50 of 1399.

k-mer=15 -cov_cutoff auto -exp_cov auto -ins_length 1000 -exp_cov 5 shows results to be n50 of 11.

My Concerns

1 :how k-mer effects the contig formation
2 : Which Files i have to see for the # of contigs formed by a particular set of parameters

[nangillala]

Hi,

1 :how k-mer effects the contig formation

Let me cite the Velvet Manual here:

As is often the case, it’s a trade-off between specificity and sensitivity. Longer kmers bring you more specificity (i.e. less spurious overlaps) but lowers coverage. . . so there’s a sweet spot to be found with time and experience.

You can read a bit more on page 16 of the Velvet Manual. There's a lot of discussion about the choice of k-mers in publications and on this forum.
If your dataset is small enough you may try to use Velvet with multiple k-mers (see page 7 in the manual).

2 : Which Files i have to see for the # of contigs formed by a particular set of parameters

You can always just use

grep -c ">" contigs.fasta

This works for all assemblies in fasta-format.

When using Velvet you get the Log file which tells you something like:

Median coverage depth = X
Final graph has X nodes and n50 of X, max X, total X, using X/X reads

[nxtgenkid10]

Hi,

Let me cite the Velvet Manual here:
You can read a bit more on page 16 of the Velvet Manual. There's a lot of discussion about the choice of k-mers in publications and on this forum.
If your dataset is small enough you may try to use Velvet with multiple k-mers (see page 7 in the manual).


You can always just use

grep -c ">" contigs.fasta

This works for all assemblies in fasta-format.

When using Velvet you get the Log file which tells you something like:

Median coverage depth = X
Final graph has X nodes and n50 of X, max X, total X, using X/X reads

What my problem is with that only

I used the k-mer to be 15 in velveth and in velvetg -cov_cutoff auto -exp_cov auto -ins_length 1000 -exp_cov 5

so as far as my understanding the contig.fa should have vaule greater than 2k in my case 30 but my files have such entries too y is it so and yu can notice as in below cov is 0.000 tooo

>NODE_9_length_21_cov_0.000000
GGGAACCGACATGAGGAACTTAGCACGGTAAACAG
>NODE_18_length_23_cov_0.000000
CCCGCGTTCAGACGCAGTTCAGACATATGGTGGACCG
>NODE_24_length_16_cov_0.000000
GGTTAAGAGCCTGCTACGCAGGCTCTTAAC
>NODE_39_length_31_cov_0.000000
TGCGCGCTGGCCAGACCGGCTTGTGCGCTGCGCAGCTGAGCGTTT
>NODE_56_length_15_cov_0.000000
CTGTCTATCAGCTGTCTCTCTTCAAGCGC
>NODE_66_length_29_cov_0.000000
CGATAACGCGCTGCCACTGACGACAGTCCAGGGTATCTCCCTC