Question about HMMER output and LTRs

[BioFalcon]

Hi,

I'm trying to run nhmmer (v3.4), trying to use a model I constructed that has a structure of an LTR. Bellow you can see how the self alignment looks like

And the alignment itself looks like this (each color regresents a different base, gaps are white):
5' End

3' End

The alignment itself is around 2.1kb. I constructed the hmm model using the command

hmmbuild --symfrac 0 --dna --informat afa --fragthresh 1 -n TEST LTR.hmm LTR.aln.fa

However, the results always falls short at 200bp (the start of the alignment). I used the following command to do the search
nhmmer --cpu 8 --tblout LTR.out --notextw --noali --seed 1992 ./LTR.hmm HMMDB_Genome

When I trim the LTR on the 3'end manually, suddenly I start getting hits that extend onto the first 200bp. Any ideas how to solve this?

Thanks

[BioFalcon]

Here are some of the files to reproduce the issue, these include the alignments used for the generation of the hmm models, the hmm models and the outputs.
ExampleFiles.tar.gz