Error with Repeat Modeler

[gprakhar]

Hello,

I am working on Repeat annotation of A. indica (Neem).
While using Repeat Modeler, it runs for 6 rounds on my scaffolds file(118093 sequences).

In the 6th round at about 14% it stops, with NO error.
On further investigation the Error log file shows an Floating Point error in the 5th round.
And the working directory contains a file called as "consensi.fa".

Now while the on line documentation for Repeat Modeler mentions the final output file to be called as "consensi.fa.classified" which is not there I have "consensi.fa".
Has the Repeat Modeler run gotten over properly?
Can I use this file for downstream repeat analysis?

Regards,
--
pg

[gprakhar]

Hello,

Should I use the "consensi.fa" file for further analysis, has the run got over properly?
Can anyone provide an insight?

--
pg

[gprakhar]

Solved

Hi,

No you can not use the "consensi.fa" file as the final result file.

I just took the top 500 scaffolds from my input file and ran Repeat Modeler on them.
It worked and produced a "consensi.fa.classified" file, no errors and the last round completed 100%.
Now trying to determine the maximum number of scaffolds that can be processed by Repeat Modeler without any error.

But now even if I split my input file into say 10 chunks, and get 10 "consensi.fa.classified", then all of them will have to be combined into a single file.
IF I simply concatenate them, them there will be many redundant sequences, hence I think the best strategy is to do a, all vs. all similarity search using something like vmatch, lastz or bastn. Then remove all sequences with more than 80% similarity and a very high e-value.

Is that a correct methodology?

Regards,
--
pg