Bioinformatics Stack Exchange is a question and answer site for researchers, developers, students, teachers, and end users interested in bioinformatics. It only takes a minute to sign up.
Sign up to join this community
I haven't been able to find the following in the documentation at http://www.repeatmasker.org/
(1) Why are certain regions classified as "Other"? Are these regions impossible to concretely classify due to alignment issues?
(2) I believe that centromeric and telomeric regions would be classified as tandem repeats. Are telomeric/centromeric regions classified as "DNA" within the hg19 version of RepeatMasker?
Are these documented somewhere?
I'll leave question 1 to someone else since I don't know.
Regarding question 2, centromeric and telomeric regions are hard masked in hg19, so there's no sequence to repeat mask.
For hg38, the centromeres are labeled ALR/Alpha with a class of centr. Telomeres are still hard-masked, but they tend to have simple repeats next to them (as annotated by repeatmasker, which makes sense).
Apparently the Other class in hg19 is comprised of SVA elements. In hg38, these are reclassified as Retrotransposon class and SVA family, which didn't exist in hg19.
repName for Other and extrapolating these must be SVA elements? You're right---these are SVA_A, SVA_B, SVA_C, SVA_D, SVA_E, SVA_F.
$\endgroup$
Oct 2, 2017 at 18:03
On (2), centromeres and telomeres primarily consist of satellite sequences. In human centromeres, the dominant type is alpha satellite (aka alphoid). GGAAT micro-satellite (aka SATIII centromeric micro-satellite) is also very common. There are less common types such as beta satellite. Because there are all kinds of centromeric/telomeric sequences of different lengths and properties, RepeatMasker doesn't put them into one class. Note that RepeatMasker has a "centr" class but it does not include all centromeric satellites (no SATIII, for example). In general, it is not advised to find centromeres with RepeatMasker output.
GRCh38 is an unusual reference genome in that its centromeric sequences are computationally generated modeled after Venter as I remember. They are mostly alpha satellite, but are not real sequences (which has always bugged me). GRC know exactly where they put these sequences. GRCh37 is a more typical reference genome. It lacks probably majority of alpha satellites, but still has many. When you map reads to GRCh37, you occasionally see ~10,000-fold read depth even though the average depth is ~30X. These regions are mostly alpha satellites. Mappers just pile all different copies of satellites onto the ones remained in GRCh37.
