What are the modification codes for various RNA modification

[Tesdhi]

Hello! I'm currently using "modkit bedmethyl tobigwig" for my Nanopore direct RNA-seq datq. For the parameter "--mod-code", I didn't know the corresponding code that can represent different modifications (m5C, m6A_DRACH, inosine_m6A, pseU) as the input for this parameter. Is there any methods to find them?

Thank you!

[ArtRand]

Hello @Tesdhi,

One way to check is to run

modkit modbam check-tags --ignore-index --num-reads 100

This is essentially the first 100 reads, it will report how many reads have modified base information and which tags. The output tables will be something like what is shown in the documentation.

You can also inspect the config.toml for the base modification models you're using. For example:

$ dorado download --model [email protected]_inosine_m6A@v1
...
$ cat rna004_130bps_hac\@v5.1.0_inosine_m6A\@v1/config.toml

[general]
creation_date = "09/13/2024, 05:33:30"
model = "conv_lstm"

[model_params]
size = 128
kmer_len = 9
num_out = 3

[modbases]
mod_bases = [ "17596", "a",] # <-- these are the modification codes this model will produce probabilities for
offset = 0
reverse_signal = true
base_start_justify = true
mod_long_names_0 = "Inosine"
mod_long_names_1 = "m6A"
chunk_context_0 = 150
chunk_context_1 = 150
kmer_context_bases_0 = 4
kmer_context_bases_1 = 4
motif = "A"
motif_offset = 0

[refinement]
refine_do_rough_rescale = 1
refine_kmer_center_idx = 3

[Yunxia-wang]

Hi,

As for this information,

mod_bases = [ "17596", "a",] # <-- these are the modification codes this model will produce probabilities for mod_long_names_0 = "Inosine" mod_long_names_1 = "m6A"

This means that "17596" is the modification code of "Inosine", and "a" is the modification code of "m6A".

I want to confirm this information. I also don't know how to distinguish the corresponding relation between codes in output file and modifications types.

[SuhasSrinivasan]

Hi @Yunxia-wang , Thanks for raising this issue!

Hi @ArtRand , It will be very helpful if the DNA and RNA modification codes are documented. Maybe here: https://nanoporetech.github.io/modkit/intro_pileup.html