Ans: Yes.

Solution 1
==========

Convert bonito_calls.subset.sorted.bam into a TSV file using modkit extract.
Then, inspect a few lines from the query_kmer column.
You will see in all of them, the third position is a C and the fourth position is a G.
This tells us that a few randomly chosen lines
contain modification probabilities only at CpGs.
You can repeat this a few times and convince yourself that all lines are like this.

```bash
mod_bam=~/nanomod_course_data/human/bonito_calls.subset.sorted.bam
output_tsv= # fill suitably
modkit extract $mod_bam $output_tsv
cat $output_tsv | shuf | head -n 10
# do the above line a few times and look at the 15th column.
# you should always see a C and a G in the third and the fourth positions
# of the query 5-mer.
```

Solution 2
==========

You are not expected to have figured this out as this involves a linux
command line tool `grep` that you may not be familiar with.

This solution build upon Solution 1.
A precise way of answering the question is to run the command below that prints
all lines which do not fit the pattern 'any base-any base-C-G-any base'.
You will find that out of the thousands of lines in the file, only 14 do not fit
this pattern: one is the header and the other 13 are k-mers like 'TACG-',
which have a central 'CG'. So, the conclusion is that all k-mers have
a C and a G in the third and the fourth positions.

```bash
input_bam=~/nanomod_course_data/human/bonito_calls.subset.sorted.bam
output_tsv= # fill with some suitable location
modkit extract $input_bam $output_tsv
awk '{print $15}' $output_tsv | grep -v -E ^[AGCT]{1}[AGCT]{1}CG[AGCT]{1}$
```