[ensembl-dev] whole chromosome alignment with LastZ

[Matthieu Muffato]

Dear Alice,

I've seen that you have sent this question to the lastz-users list and 
Bob replied with some tips.

We don't get this error we get because we have engineered a workflow 
around LastZ that does what Bob suggests, i.e. the masking, the 
splitting with overlap and its post-processing, and some chaining.

It's not easily usable by external people, though, because it heavily 
relies on an infrastructure that's built in Ensembl. But Bob gave some 
options to run those steps by yourself, which should address the issue.

Best,
Matthieu

On 25/03/2020 11:36, Alice Iob wrote:
> Good morning,
>
> I am a Phd student woking on plant genomics and I am stuggling with an 
> issue regarding LastZ.
> I choose to use LastZ because it was used for plant genome alignments 
> in Ensembl Plants, so I hope that someone who succesfully used it 
> before can help me with this.
> I am trying to align two references genomes from very close species: I 
> have two FASTA files,
> representing the same chromosome in the two species, each around 800Mb 
> long, with at least one long repetitive region.
>
> the command I am using:
>
> lastz target.fasta query.fasta --notransition --step=20 
> --maxwordcount=70 ‑‑exact=20 --chain --gapped --ambiguous=iupac 
> --rdotplot=plot --format=differences > alignment.differences
>
> I always get the same error:
>
> FAILURE: in add_segment()
> table size (4,869,542,152 for 101,448,794 segments) exceeds allocation 
> limit of 4,294,967,279;
> consider raising scoring threshold (--hspthresh or --exact) or 
> breaking your target sequence into smaller pieces.
>
> I tried several strategies to overcome this issue:
> increasing values of exact up to 100
> using values of hspthresh up to 10 000
> adding --seed=match12
> dividing my target sequence in two (one multiFASTA file)
> working with just half chromosome (400Mb)
> set the parameters as they were set to align /T. aestivum/ and /A. 
> tauschii/ (https://plants.ensembl.org/mlss.html?mlss=9814)
>
> Still, I get the same error.
>
> Just a few times I was able to get an output (e.g. when exact=100), 
> but it is always more than 700Gb big, thus, even if the file is 
> generated, I run out of memory and I can not work on it.
>
> I also used LastZ_32 but the process gets killed without giving me any 
> info.
>
> I was wondering if you can help me with this issue, maybe I am not 
> using properly some of the options, or give me some advice on how to 
> proberly deal with this alignment.
>
> Thank you.
>
> Alice Iob
> PhD student
> Plant and Animal Genomics Program
> CRAG, Centre for Research in Agricultural Genomics
> Campus UAB - CRAG Building | 08193 Cerdanyola | BARCELONA
> Office: 3.01
> Tel. +34 935636600 ext 3351
>
> _______________________________________________
> Dev mailing list    Dev at ensembl.org
> Posting guidelines and subscribe/unsubscribe info: https://lists.ensembl.org/mailman/listinfo/dev_ensembl.org
> Ensembl Blog: http://www.ensembl.info/

-- 
Matthieu Muffato, Ph.D.
Ensembl Compara Principal Developer
European Bioinformatics Institute (EMBL-EBI)
European Molecular Biology Laboratory
Wellcome Trust Genome Campus, Hinxton
Cambridge, CB10 1SD, United Kingdom
Room  A3-123
Phone + 44 (0) 1223 49 4631
Fax   + 44 (0) 1223 49 4468

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ensembl.org/pipermail/dev_ensembl.org/attachments/20200326/6a806e62/attachment.html>