Query about special tRNAs

[iqbal-lab]

Got this query on LinkedIn from Igor Tolstoy

technical question about search sensitivity: is it possible to find all trnas with very rare anticodon mutation, just one nucleotide change in wobbling position frequency expectation : 1-2%

I asked for further info, and he said

https://pmc.ncbi.nlm.nih.gov/articles/PMC8007984/

query very special - specific three letters in anticodon positions

according publication prokaryots do not have 16 anticodons out of 64
But from my partial analysis they all presented with very low frequency.
How to find them on global scale 2M genomes.
Interesting estimate cost

[igortru]

Prokaryotic tRNA molecules are typically about 73 to 93 nucleotides in length. Most commonly, they are around 76 nucleotides long
from 20 to 100 per genome
best current way identify trna and anticidon - trnascan-se
but you need retrieve all genomes, annotate them
and then search mutation

technically , it is just test-case.
same method can be applied to any gene where we need find specific SNP

[shenwei356]

We've done a similar search before, querying for a special SNP of a Mycobacterium tuberculosis gene.

1. Searching with the query (with or without SNP), with -a added to output sequence as well.
2. Filtering with a short (13 bp) sequence containing the SNP and up- and downstream 6 bases.

MTB genes are very conservative, but tRNA genes might be less conservative, especially for those of different species.

Besides, considering the length of tRNAs (73 to 93 bp), a new LexicMap index must be built with denser seeds, by decreasing the value of -D/--seed-max-desert from the default 100 to 70 or 50.

All ATB genomes have been annotated with Bakta, so retrieving genes and checking mutations can also be feasible.