[khmer] cannot run the latest version of normalize-by-median.py

C. Titus Brown ctb at msu.edu
Tue Aug 27 18:55:52 PDT 2013

On Wed, Aug 28, 2013 at 08:03:57AM +0800, cy_jiang wrote:
> Hi Titus,
> I have downloaded the latest version of khmer and tried to run the normalize-by-median.py (command line:/khmer/scripts/normalize-by-median.py -k 20 -C 20 -x 7.5e9 -N 4 --savehash mh.kh T_allR1R2.py), but it prompted up with the following information:
> Traceback (most recent call last):
>   File "/home/work/khmer/scripts/normalize-by-median.py", line 241, in <module>
>     main()
>   File "/home/work/khmer/scripts/normalize-by-median.py", line 140, in main
>     ' - kmer size =    {:d} \t\t(-k)'.format(args.ksize)
> ValueError: zero length field name in format
> This command line ran pretty well with the older version of khmer. Can you give me a clue what is going on there?

That's weird... you're setting k properly on the command line... hmm.

What does 'python -V' return?

> Besides, I wonder how to choose a better k-mer size. As I am going to assemble my dataset with Trinity, I am thinking would it be better using a k-mer value of 25 since Trinity set this default k-mer size when it does the assembly. 

We haven't found it to matter much; while the k-mer size is similar in concept
to the k used by de Bruijn graph assemblers, when you look a bit deeper it
is almost completely unrelated in practice.  I can tell you that a k of 20
seems to do fine for Trinity, but I can't tell you what a k of 25 would do.


More information about the khmer mailing list