search

wyang17 / SQuIRE

Software for Quantifying Interspersed Repeat Expression
Other
49 stars 29 forks source link

Help on strand L1 expression #48

Open GuillermoPeris opened 3 years ago

GuillermoPeris commented 3 years ago

Hi,

I am beginning to work with Squire to analyze L1 mobile elements in knockout cell lines. I am specifically interested in obtaining antisense expression on L1 (that should include transcripts from L1 antisense promoter and "run on" transcripts from nearby gene promoters).

I have some trouble in getting that information from strand output (tx_strand and TE_strand). According to SQuIRE Count output explanation, TE_strand is orientation of TE insertion and tx_strand the strand of TE transcription. My guess is that I can get antisense L1 transcripts by choosing counts where TE_strand is different from tx_strand (TE_strand != tx_strand). Am I right?

Thank you for your support and for this useful software,

Guille.

rpg18 commented 3 years ago

Hi @GuillermoPeris !

As far as I know, that information is also included in the TE locus name in the first column of the counting tables, such as: chr10|48318854|48319078|L2:L2:LINE|326|-,+ or chr7|94986341|94986764|L1MA5:L1:LINE|133|+,+

For the anti-sense expression of L1, you can easily extract "-,+" or "+,-" entries.

I am performing the same approach as you. Let me know if you would like to further discuss about L1ASPs!

All the best, Raquel

GuillermoPeris commented 3 years ago

Hi Rachel!

Thank you for info confirming my hypothesis. Concerning L1ASP, the problem with SQuIRE is that cannot distinguish (as far as I know) reads from L1 antisense promoter and passive antisense transcription. I suggest you should try package L1EM (https://github.com/FenyoLab/L1EM). It really works to find specifically L1ASP.

Good luck!! Guille