[jvanleuven]

well, there are probably at least two genomes in there and lots of contaminating host DNA. One genome is pretty easy because it has >95% sequence identity to an annotated genome. The other is more of a challenge b/c is likely quite diverged from anything else. both are under 500kb in length.

jt

also i tried aggressive filtering. reads >50bp, quality >30 for 100% of the read.
cut it down to 5 million PE reads. didn't help. did lower the coverage of the more abundant genome accordingly, but also reduced the contig lengths and didn't effect the number of Ns in the contigs.

[aloliveira]

cliffbeall

I would try being very aggressive in quality filtering/trimming. Most accounts that I have seen don't see improvement beyond 50X coverage, so you should be able to be selective with your input data.

[filter using fastx-tool kit - remove reads < 20bp in length, <30 quality over 90% of the read

Is it a good quality filtering/trimming no?

[jvanleuven]

Not sure if that is good enough filtering, this is the first time i've tried it.

I think it is pretty good. The first few bases and the last few bases have average qualities around 34, the rest are higher.

[jvanleuven]

Thanks for all the help. Still not sure about how to fix the N gaps created by scaffolding. I am thinking about using Cap3 to try and cover the gaps with contigs created from a single read assembly using all the quality filtered reads.

What software may be used to map lots of shorter contigs to longer contigs to give a consensus sequence?

[nickloman]

It's not really a case of fixing them as Velvet is simply telling you that it thinks two contigs are joined together with a gap of a specific length by reference to paired-end information. But it cannot fill in that gap with sequence because there is no apparent unambiguous path through the assembly graph. This is usually a result of repetitive sequence longer than the read length, but sometimes could be due to low coverage.

Basically you have scaffolds formed of contigs (actually I wish Velvet would call the output scaffolds.fa not contigs.fa to make this more clear, or produce an AGP file, but that's another matter).

You might find that GapCloser (part of SOAPdenovo) can be coaxed into filling some of these gaps, by performing local assemblies using pair-halves which fall within the gap region as determined by some mapping approach (http://soap.genomics.org.cn/about.html). IMAGE is another utility that does this (http://genomebiology.com/2010/11/4/R41). In my experience results are mixed and it will depend on the repeats in your genome.

[nickloman]

And, PS

Read filtering is probably not the significant factor here.

[jvanleuven]

yeah, i think my reads are sufficiently filtered. I think my problem may be a result of such widely varying coverages, or maybe variance in my insert size, or possibly lots of repeats? both genomes are AT rich (25%).

if I can resolve some of the N gaps with single reads, my contigs are long enough.

jt

[nickloman]

99 times out of 100 it will be due to repeats. It's normal!

[jvanleuven]

Yeah, but a lot of my contigs look like this;


NODE_8445_length_3509_cov_38.863495
ATGGGATTGGAATACAAGAAGTTACTTCAAATATATAACATGTCGAACAACAAACATGGA
GATAATCGATATCGCTAAATCAAAGCTATCAGAAAGACATCCTGAATGTTGATTTTAAAA
CAATTGATAACGAAGTAATATTGGAATTGATATTCGATTTATCCAATGGTATTAAAATAA
ATGTTCAAGACATAGATATAAAGAAGGATATTAAGTTCTTAGAATCACGTATTGATCTAA
AAGATGATAAATATCAATATCAATTATACGATGGGATAACAGGTTTAAAACATGATAGAA
AGACAACGGTTGGAATTTATGTATGTTTATAAATTAAATCACATGGTAGATGATAAAATG
TATGCCAGATCTATAGGACCTTATTCAACTATTACACAACAACCACTTAAAGGTAAGTTC
CATAAAGGAGGACAGATGTTGGGTGAGATGGAGGTATGGTTGTTGTAAAGCTATGGTGTT
GCATTCGTTACAAACGAAGCATTGACCGCTAAATGTGATGATATTCAAGCATGAAACAAA
TTGCATGATAACATGTTATTCGGTACACCATTGNNNNNNNNNNTTTAGATTAGTAATGTT
GTATGGCTTTAGATTAGTAATGTTGTATGGTGATTCGATTTTGTTATTAAAATATGATGT
TAATTTATGATTATTGATTATAAAGAATAGCGACAAGTTGTAATATTATTGGTATGANNN
NNNNNNNGAGTTAACGATAGCCTCTCCAGAGGCTACAATGACCTATTCTTGGGGAGAGGT
ATACAACGGTATTGGGTTCGATTTTGTTGTTAAAATATGATGTTAATTTGTGATTATTGA
TTATAAAGAATAGTGACAAGTTGTAATATTATTGGTATGAAAGTTATCAAGTAGTATATT
GTCTGTTACTAATATAATCAATGAGATATATAAACACTGTGATTTGTTAAAGGTANNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNAATTATCAACATAAGTTAATAGGTTGATTGATAAAACAACAAAT
TGATAGATTATTGGGTATAATTTGGTCAAANNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNAATTT
GAATTATTTACTCTAAATTGTATAAATCACTTGGTAATGATGGATTCAATAAATCTCTAA
GATAACAAACGGTTTTTATTTCTTAAATAACTATCTGATAACAAAACAGTTGCTTTAGTG
ATATCGTTATCAATTCTTTTCAGAAAACCTAATCGGTTGTCTTCACAAATTCCTTTAACT
AATTTAATCTTATTTATAAATTCCTTTCTAACTGTTCTGATACTTGTTTTATATCTTTCA
AACATTTTAGATAATTCATTCAAAGCAATCTTTTGTCTCTGTGTTATTATTGATGGTAAG
GCTTAACTTGATATGATTCCCGTTAAGTNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNAATTATATCGTTTGTTCTAAGTTTACA
TATTGGTCGTAGTTCAATAATTGTATCAGAAAGATAAGTACATTNNNNNNNNNNGGACTT
CAATGATTGATTTTTTGTTGTTCTTATTTCATTAACAAGAGACATATCAAGTCTTGATGG
TAAAATTTTGTTTGATATAAGATCAAAATTTTTAATTTAAGTTTAATAATAATTGGTTCA
ATTGTATGTATATATCTTCCAATGTTAACTTTGAATAATTTGGCTTTAATTTAAACAATT
ATAAGAATTAGTTCAGTTCATCTAATATTCAGATGGAAATTAAACGAGAAATAACATATT
GAATTATAGAACAAACAATAAGAGAACCAACAATAAGATAATTTGGAATTGTAATCCAAT
TTAGATGAGTTTAGTTTATAAATTTTGATAATCTCTGGAATATATTAATTTAAATGCAGT
TGTTGGTAGAATCTNNNNNNNNNNGAAAATAATATATTGAATTATAGAATCAACAATAAG
ATAATTTGGGATTATAATCCAATTTAGATGAGTTTTAGTTATAAATTTTGATAATCTTTG
GAAGATATTAATTTAAATGCAGTTAGTTGGTAGAATTTGGAAATAAACAGATGATCTTTA
AATCACTATTTTTCTTAAAANNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNAATAAAATATTAACAGTTGATCAATAA
TTCGATATTATTAAACGAATCCAATGAAAATAAACAAGTTAATAACNNNNNNNNNNGATA
AAATATTAACAGTTGATCAATAATTCAATATTATTGAATGAATCCAATGAAAATAAACAA
GTTAATAACCAATTTTATTGGCATTGATAGATAGTTGGTTTCTATTCCTAACAAATACTT
GATTACTTAATAAGATTCAAAGGAGAGATAAATAACTTATGGTCCGTTTTAAACTAATTA
ATGATTNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNGACAATTAATAACAATTTATATCTGAATTATATTCGATTATTTTAGAAA
GAGATCGTTTAAAAAAAAANNNNNNNNNNTATTGATAGATAGTTGGTTCTTATTTCTAAC
AAATATTTGATTAGTTAATAAGTTTCAAAGGAGAGTATAATAACTTANNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNGATTATTTTCTCTATGT
TTATATTTATAATTTTCGATGAAATCATATTCTTGATAACTCTCTAATAAGAGTAATTTT
TAAATCTAATAATATTAAAACTCAGTTTGAGAAATTAGTTTGTTACGATCGAATATAAAT
TAAAATTAAGCACAACAAACCTAATAATTTTGATCAAAATTTATTAAGTGTTTAATTTAT
GTGTGTTAAATATTATAAATAATACTTAATTAACTTAATTTAAAAATTAATTTGTTGGTA
ATAATTTTTATTTTTTAATGTGTTAAGATTTATCGTATTAATTATATAATAATTTATTAT
TCTTAATAGAAGTAACACTAATAACTTATCGGATATTGATTGAATCTATTGAAATTGATG
TTNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNAGCGGTTACATATATTATTTTTGAA
TTGTGAGAGATATGAATTTCATTTTCTTTATTCAAAATAAACAATTATTTATTTCAATT