contig overlaps in velvet 1.2.03

[hard998]

Hi, I did a denovo assembly using velvet 1.2.03 which gave me 1626 contigs,

However, when I examined the contig.fa output by velvet, eg
>NODE_1331_length_2655_cov_63.920151
>NODE_1395_length_1978_cov_42.813953
...

I found 689 contigs have tail-to-tail matches with each other by doing BLAST, such as
NODE_1331_length_2655_cov_63.920151: 2656 tgcggctggtcggctcgtcgcaggcgaacatgttcgcctccatc 2699
||||||||||||||||||||||||||||||||||||||||||||
NODE_1395_length_1978_cov_42.813953: 2022 tgcggctggtcggctcgtcgcaggcgaacatgttcgcctccatc 1979

Can anyone tell me why didn't velvet merge those overlaps? Is that correct for me to join contigs into larger contigs? Thank you!

[seb567]

Originally Posted by hard998

Hi, I did a denovo assembly using velvet 1.2.03 which gave me 1626 contigs,

However, when I examined the contig.fa output by velvet, eg
>NODE_1331_length_2655_cov_63.920151
>NODE_1395_length_1978_cov_42.813953
...

I found 689 contigs have tail-to-tail matches with each other by doing BLAST, such as
NODE_1331_length_2655_cov_63.920151: 2656 tgcggctggtcggctcgtcgcaggcgaacatgttcgcctccatc 2699
||||||||||||||||||||||||||||||||||||||||||||
NODE_1395_length_1978_cov_42.813953: 2022 tgcggctggtcggctcgtcgcaggcgaacatgttcgcctccatc 1979

Can anyone tell me why didn't velvet merge those overlaps? Is that correct for me to join contigs into larger contigs? Thank you!

These are called repeats. You probably have more than 2 contigs with this sequence on one end. Therefore, Velvet did not merge them because that would cause an assembly error, which produce chimeric contigs.