SOAPdenovo2 .config settings

[jpummil]

So, I recently gave SOAPdenovo2 a try on a large genome assembly which I'd previously been using Velvet and Ray on. I know SOAP was used with great success on te Giant Panda among many others. One setting of question for me right now is the global "maximal read length". I have a number of both PE and MP sets from HiSeq that are 101bp long. But, I also have a set of PE from MiSeq that is 267bp long. It would seem that you would do as instructed and assign max_rd_len to be the 267 number, but does that have an adverse affect on all of the 101bp stuff?


#maximal read length
max_rd_len=267

Couldn't find an example where you set this within the [LIB] parameters for each individual set of data...

[jpummil]

Hahaha! 134 views and no responses. OK, I have set up a run with the parameter set inside the [LIB] areas for each read type...we'll see if results are different than with it set globally...STAY TUNED!

[vivek_]

That seems the best course of action. When I worked with SOAPdenovo or for that matter any de novo assembler, I was advised to generate multiple assemblies with different parameters and retain the one with the best metrics in the end.

[jpummil]

So, answer is...YUP, it makes a difference!

Quast results when setting max_rd_len globally and using longest read length for value:

Assembly horridus.contig
# contigs (>= 0 bp) 24780968
# contigs (>= 1000 bp) 215761
Total length (>= 0 bp) 3052603782
Total length (>= 1000 bp) 312510366
# contigs 761946
Largest contig 9207
Total length 691019446
GC (%) 37.43
N50 938
N75 687
L50 249843
L75 466230
# N's per 100 kbp 0.00

Quast results when setting max_rd_len per [LIB] using specific read length for value:

Assembly horridus.contig
# contigs (>= 0 bp) 15252819
# contigs (>= 1000 bp) 329326
Total length (>= 0 bp) 2030747645
Total length (>= 1000 bp) 613954195
# contigs 695060
Largest contig 15596
Total length 873677890
GC (%) 37.08
N50 1461
N75 910
L50 182160
L75 372629
# N's per 100 kbp 0.00