support for PacBio reads?

[Carlos Cano]

Hello, 

I wonder whether MOSAIK can handle PacBio reads, although this technology is not explicitly supported (not currently listed as an option for the -st parameter in MosaikBuild). If not, are you planning to include support for PacBio in future releases? 

 Thanks, 

 -Carlos  

[Wan-Ping Lee]

Hi Carlos,

I indeed tested MOSAIK on PacBio V. cholerae reads. The program is running; however, I need to employ other parameters rather than the default ones. Otherwise, the aligned rate is very low. The parameter set “-hs 10 -mmp 0.5 -act 15” was used as opposed to the default values “-hs 15 -mmp 0.15 -act 55”.

Best,

Wan-Ping

 -Carlos  

--
You received this message because you are subscribed to the Google Groups "mosaik-aligner" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mosaik-aligne...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
Wan-Ping Lee
Postdoctoral Research Associate
Department of Biology, Boston College

Tel: +1 (617) 552-2922
wanpi...@bc.edu

[Wan-Ping Lee]

It seems that I forgot to answer the question about MosaikBuild. I used "-st 454" for PacBio datasets.

Best,

Wan-Ping

On Sat, Feb 9, 2013 at 6:49 AM, Carlos Cano <carloscan...@gmail.com> wrote:

 -Carlos  

--

You received this message because you are subscribed to the Google Groups "mosaik-aligner" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mosaik-aligne...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Carmen Navarro]

Hi,

I have been trying this configuration you mentioned and the results are good. However, I wonder whether there is a way to improve the execution speed (i.e. some efficiency-oriented parameter configuration). For shorter reads Mosaik is very fast. I have tried to run the software with several processors, but I get the same execution time for 1, 2, 4, 8, 12 processors. Maybe the parallelization is not compatible with PacBio reads' length/error rate?

Thanks in advance.

Carmen

[Wan-Ping Lee]

Hi,

I knew the performance of this parameter setting is very slow. I got the same issue here. I think the reason is -act 15 which means any hash region larger than 15bp will be enable a Smith-Waterman (SW) for it. You may try larger number for -act to speed up.

Ongoing development replaces the current banded SW by a SIMD SW. The initial test shows two-fold speedup. However, I haven't finished and released it.

Best,

Wan-Ping

[Huang Ke]

Hi Wan-Ping,

I wanted to use Mosaik to do the alignment for the PacBio reads. I'm exited to have found this post but i came across some issue when I run MosaikAligner using the command "MosaikAligner -p 7 -in pacbio_4571661.mkb -out pacbio_4571661.mka -ia genome.dat -j genome_10 -hs 10 -mmp 0.5 -act 15 -annpe 2.1.78.pe.ann -annse 2.1.78.se.ann",  an error message jumped out:

Mosaik [1;31mAligner [0m 2.2.26                                                2014-03-28

Wan-Ping Lee & Michael Stromberg  Marth Lab, Boston College Biology Department

------------------------------------------------------------------------------

- Using the following alignment algorithm: all positions

- Using the following alignment mode: aligning reads to all possible locations

- Using a maximum mismatch percent threshold of 0.5

- Using a hash size of 10

- Using 7 processors

- Using a Smith-Waterman bandwidth of 3265

- Using an alignment candidate threshold of 15bp.

- Setting hash position threshold to 200

- Using a jump database for hashing. Storing keys & positions in memory.

- Using a homo-polymer gap open penalty of 4

- loading reference sequence... finished.

- loading jump key database into memory... finished.

- loading jump positions database into memory... finished.

[1;33m

Aligning read library (40908):

[0m

 0% [ [1;32m                                     [0m]                                  |

17% [ [1;32m=====>                               [0m]                                  /

17% [ [1;32m=====>                               [0m]                                  -

17% [ [1;32m=====>                               [0m]                                  \

17% [ [1;32m=====>                               [0m]                                  |

17% [ [1;32m=====>                               [0m]                                  /

17% [ [1;32m=====>                               [0m]   2,294.5 reads/s       ETA 14 s -

17% [ [1;32m=====>                               [0m]   2,294.5 reads/s       ETA 14 s \

17% [ [1;32m=====>                               [0m]   2,294.5 reads/s       ETA 14 s |

17% [ [1;32m=====>                               [0m]   1,502.4 reads/s       ETA 22 s /

17% [ [1;32m=====>                               [0m]   1,502.4 reads/s       ETA 22 s -

17% [ [1;32m=====>                               [0m]   1,502.4 reads/s       ETA 22 s \

17% [ [1;32m=====>                               [0m]   1,105.6 reads/s       ETA 30 s |

17% [ [1;32m=====>                               [0m]   1,105.6 reads/s       ETA 30 s /

17% [ [1;32m=====>                               [0m]     867.0 reads/s       ETA 39 s -

17% [ [1;32m=====>                               [0m]     867.0 reads/s       ETA 39 s \

17% [ [1;32m=====>                               [0m]     867.0 reads/s       ETA 39 s |

17% [ [1;32m=====>                               [0m]     707.3 reads/s       ETA 47 s /

17% [ [1;32m=====>                               [0m]     707.3 reads/s       ETA 47 s -

17% [ [1;32m=====>                               [0m]     707.3 reads/s       ETA 47 s \

17% [ [1;32m=====>                               [0m]     592.8 reads/s       ETA 57 s |Alignment score and position are not consensus.

Do you have any clue about this error?  is there any pre-process step I need to take before running the program?

Thanks

Ke