High LTR unknown percentage

[rajatpruthi4]

Hi Shujun

Thanks for creating EDTA, keep doing the great work. I wanted to share some observations from my analysis. I have been running EDTA on several red raspberry assemblies and have noticed an unusually high percentage of LTR unknown repeats. Interestingly, most of these repeats seem to be concentrated in the centromeric regions. This trend is consistent across more than 100 genome assemblies.

Could you please provide your insights on why I am observing this unusually high percentage of LTR unknown repeats? I have attached the .sum output and TE density plot to this issue for your reference.
EDTA.TEanno.density_plots.pdf

<style> </style>

Repeat Classes
==============
Total Sequences: 11
Total Length: 287904378	bp
Class	Count	bpMasked	%masked
=====	=====	========	=======
LINE	--	--	--
I	483	427336	0.15%
L1	5010	2376953	0.83%
LTR	--	--	--
Copia	11758	11801425	4.10%
Gypsy	14344	20919798	7.27%
unknown	86279	67771835	23.54%
SINE	--	--	--
tRNA	382	40106	0.01%
TIR	--	--	--
CACTA	9395	4295459	1.49%
Mutator	19391	5551703	1.93%
PIF_Harbinger	14354	4224540	1.47%
Tc1_Mariner	343	118358	0.04%
hAT	12807	4413702	1.53%
nonLTR	--	--	--
pararetrovirus	27	24109	0.01%
nonTIR	--	--	--
helitron	34547	14855621	5.16%
repeat_fragment	19072	5125036	1.78%
total interspersed	228192	141945981	49.30%
snRNA	78	8331	0.00%
Total	228270	141954312	49.31%

[oushujun]

Thanks for sharing your results. One reason is that these LTR/unknown are misannotated from centromeric repeats. You will need to curate them to make sure. Highly recommend using TEtrimmer for this purpose. You may read the paper for more information about the tool usage and intepretation. Another reason - they are real LTR elements with unknown classifications - again, curation will tell you more.

Thanks!
Shujun