How should I approach this assembly?

[OstanNick]

Hello,

I am trying to de novo assemble a gene locus which contains repetitive DNA in the form of

i.e. CTTCTTCTTCTTCTTCTT

along with repeats that are 10kb in length.

I have the sequences of 8 BACs containing our insert. They were sequenced with PacBio so I have the CCS reads, the longest reads and the filtered reads.

I have been reading papers on what the best software may be to assemble these things but I'm very new in this field and definitely need some assistance regarding the overall process.

As of now, I understand I will need to

a) Decontaminate my sequence (theres a lot of E.Coli stuff in the sequences) while retaining the vector sequence. So I used/am going to use DeconSeq which seems to work pretty well.

b) Correct the long reads using the CCS reads and PacBioToCA. My boss is afraid that since the locus is very repetitive, this may generate false reads.

c) Assemble these reads into contigs
Which is the best software to use?

d) Assemble the contigs into scaffolds

Is this correct?

Thanks for any help,
Nick

[flxlex]

How much coverage do you have of each PacBio read type?

a) yes

b) repeats of this length are a problem to correct. If you have enough PacBio data, 60-100x, you can use HGAP, i.e. correcting the longest raw reads with all raw reads, assembly of the corrected reads. Otherwise I agree with your boss that the repeats may not be corrected correctly.

c) celera or newbler (version 3)

d) only works if you have additional long-range information, i.e. mate pairs. Use BESST, Bambus or the like (see the recent paper benchmarking most scaffolders). Or use all reads (pacbio + matepair) in celera/newbler.