empty tmpConsensi.fa

[MikiSchikora]

Hello,

I am trying to run RepeatModeler 2.0.1 (installed with conda) for two different databases with the following command:

RepeatModeler -database <db_name> -pa 4 -LTRStruct -debug , where <db_name> can be either chr6.fasta (just one chromosome) or complete_genome.fasta (the whole genome).

RepeatModeler works well for the whole genome, but fails for chr6.fasta displays the following message:

RepeatClassifier Version 2.0.1
======================================
Search Engine = rmblast
  - Looking for Simple and Low Complexity sequences..
FastaDB::compact - Error could not locate file tmpConsensi.fa.masked!
 at /data/anaconda3_cluster/anaconda3/envs/perSVade_env/share/RepeatModeler/RepeatClassifier line 333.
Classification Time: 00:00:00 (hh:mm:ss) Elapsed Time

In fact, the files tmpConsensi.fa.masked and tmpConsensi.fa exist, but they are empty, which probably raises this error. In addition, the final chr6.fasta-families.fa is not created. However, the files consensi.fa and combined.fa exist and contain repeat families, suggesting that I should be getting a non-empty output.

I wonder whether 1) this is a bug in the program, or 2) I just have to run RepeatClassifier on the families found.

If 2), should I run RepeatClassifier on consensi.fa or combined.fa?

Would this be a solution to any situation where there are errors in RepeatClassifier due to empty 'tmpConsensi.fa'?

I also attach the full log for clarity:

chr6_repeatModeler_out.txt

Thanks for your time.

[jebrosen]

In fact, the files tmpConsensi.fa.masked and tmpConsensi.fa exist, but they are empty, which probably raises this error.

That does seem wrong; tmpConsensi.fa is supposed to simply be a copy of one of the consensus files. What happens if you re-run RepeatClassifier manually?

However, the files consensi.fa and combined.fa exist and contain repeat families, suggesting that I should be getting a non-empty output.
should I run RepeatClassifier on consensi.fa or combined.fa?

How many families are in each? Since there are only a few files, it should be easy to run RepeatClassifier on both and compare.

[leon945945]

Hi, the RepeatModeler program almost done, but bug was arised as following:

something went wrong with the TRFMask program. The tmpConsensi.fa.masked file was missing or empty!

The consensi.fa.classified file was not generated, what's wrong with the TRFMask program and how can i fix it. Thanks!

[jebrosen]

@leon945945 that error message is from an older version of RepeatModeler; in fact, that exact part of the code has been changed as part of some cleanup and fixing on RepeatClassifier last year. I would first try downloading the latest RepeatModeler in a new directory and running the new version of RepeatClassifier on your consensi.fa file.