Does Falcon work with .bas.h5 and .bax.h5 files?

[JohnYu-yuzhongyi]

Hi,

I'm a new to bioinformatics so apologise if my question is stupid.

I'm trying to use falcon to assemble my PacBio reads. In the 'input.fofn' file, should I put all my .bax.h5 and .bas.h5 files?(I have 2 .bas.h5 files and 6 .bax.h5 files) Or should I convert them into .fasta files before I use falcon?

I tried to run falcon with those .h5 files but an error occurred.

[INFO](local) '/home/zhongyi/pacbio/0-rawreads/prepare_rdb.sh'
[WARNING]Call 'bash /home/zhongyi/pacbio/0-rawreads/prepare_rdb.sh 1> /home/zhongyi/pacbio/0-rawreads/prepare_rdb.sh.log 2>&1' returned 256.
[ERROR]Contents of '/home/zhongyi/pacbio/0-rawreads/prepare_rdb.sh.log':
trap 'touch /home/zhongyi/pacbio/0-rawreads/rdb_build_done.exit' EXIT
+ trap 'touch /home/zhongyi/pacbio/0-rawreads/rdb_build_done.exit' EXIT
cd /home/zhongyi/pacbio/0-rawreads
+ cd /home/zhongyi/pacbio/0-rawreads
hostname
+ hostname
vinculin
date
+ date
Tue Oct 27 16:38:55 GMT 2015
fasta2DB -v raw_reads -f/home/zhongyi/pacbio/0-rawreads/input.fofn
+ fasta2DB -v raw_reads -f/home/zhongyi/pacbio/0-rawreads/input.fofn
fasta2DB: Cannot open /home/zhongyi/pacbio/data/1_p0.bas.h5.fasta for 'r'

I noticed a program called fasta2DB was trying to open a .bas.h5.fasta file which doesn't exist. I wonder does that mean the program assumes the input files should all be fasta files?

[mseetin]

Hi John,

You need to run an RS_Subreads job either in SMRT Portal or SMRT Pipe to generate subreads.fasta files for each of your SMRT cells. Then list the full paths to those in your fofn.

[JohnYu-yuzhongyi]

Thanks @mseetin !

But if I only use fasta files, does that mean I'm losing all the other information like sequence quality etc.?

[mseetin]

Hi John,

Falcon doesn't use that information. That information isn't as useful for the overall assembly as it is for the final sequence polishing with Quiver. Falcon doesn't perform the polishing step. Instead, what you'll need to do is concatenate the p_ctg.fa and a_ctg.fa files you get from Falcon into one fasta file, upload that as a reference into SMRT Analysis, and then run a BAM_Resequencing job using your SMRT cells as input and your new reference as the reference. This will produce a polished consensus sequence that is approximately QV 50. This takes as input the full .bax.h5 files and uses all the sequence quality information to produce it.

[JohnYu-yuzhongyi]

Thank you so much @mseetin! I'll give it a try

[pb-cdunn]

Linked from FAQ. Thanks, @mseetin.