error in hmmufotu-build

[chassenr]

Hi,
I would like to use hmmufotu with the Silva database. I tried running the hmmufotu-build command using the silva aligned sequences and the tree file that are available for qiime2. However, I get the following error with hmmufotu-build rep_set_aligned/99/99_alignment.fna trees/99/99_otus.tre -n Silva132_hmmufotu_GTR -s GTR -p 20:

hmmufotu-build: MSA.cpp:156: long int EGriceLab::HmmUFOtu::MSA::loadMSAFasta(const EGriceLab::HmmUFOtu::DegenAlphabet*, std::istream&): Assertion `concatMSA.length() == numSeq * csLen' failed.

Do you know what causes the error? Am I using the wrong kind of input files? Are you by any chance planning to also put pre-built databases for Silva on your website?

Thanks a lot!

Cheers,
Christiane

[e00011027]

Hi Christiane, Thank you for trying our tool HmmUFOtu. Please be advised that as described in the original article, HmmUFOtu requires fully evaluated phylogenetic trees with all branch length pre-estimated, and every and all nodes in the aligned sequences being present in the tree. Until last time we checked, Silva trees do not satisfy these criteria: the alignment contains more sequences than that present in the trees, in which are just a manually curated "reresentative" seqs in the alignment; also Silva trees are guide-trees only -- no branch length was estimated. Please double confirm all these criteria with you Silva database input files, and let us know the results. PS: We are also actively working on advanced version of HmmUFOtu that may handle guide-trees only in the future. Best regards Qi

…

On Mon, Dec 10, 2018 at 9:13 AM Christiane Hassenrück < ***@***.***> wrote: Hi, I would like to use hmmufotu with the Silva database. I tried running the hmmufotu-build command using the silva aligned sequences and the tree file that are available for qiime2 <https://www.arb-silva.de/fileadmin/silva_databases/qiime/Silva_132_release.zip>. However, I get the following error with hmmufotu-build rep_set_aligned/99/99_alignment.fna trees/99/99_otus.tre -n Silva132_hmmufotu_GTR -s GTR -p 20: hmmufotu-build: MSA.cpp:156: long int EGriceLab::HmmUFOtu::MSA::loadMSAFasta(const EGriceLab::HmmUFOtu::DegenAlphabet*, std::istream&): Assertion `concatMSA.length() == numSeq * csLen' failed. Do you know what causes the error? Am I using the wrong kind of input files? Are you by any chance planning to also put pre-built databases for Silva on your website? Thanks a lot! Cheers, Christiane — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AG-ecQ2aA7ElAUSFQ_mjwqquhx3fU6ojks5u3mvsgaJpZM4ZLccE> .

[chassenr]

Hi Qi,
thanks a lot for your advice. I also contacted the silva team and they confirmed that in its current state, the silva tree may not be suitable for hmmufotu - at least not with some extensive validation. The problem of having different numbers of sequences in the alignment and the tree shold be much easier to fix by some file conversions in e.g. arb, though.