Marcus Stoiber

Here is a [link to the docs](https://nanoporetech.github.io/tombo/text_output.html) for the text output commands. From there you can also find a [link to the wiggle format here](https://genome.ucsc.edu/goldenpath/help/wiggle.html). If this does not provide...

Defining methylated sites versus un-methylated sites is a bit complicated as methylation often occurs at a fractional level. This is why tombo outputs an estimated fraction of modified bases at...

The fraction of methylated bases estimated by tombo at each reference position is found within the `dampened_fraction` wiggle file. As defined in the previously attached link, each line in a...

These results certainly suggest a high error rate for 6mA detection. The all-context 6mA model generally has much higher error rate than the provided `dam` model. A more in depth...

Ah, I take back the non-existent feature for per-read roc curves. The command that may help you is the `tombo plot sample_compare_per_read_roc` command. Passing this command the per-read statistic files...

The `tombo plot sample_compare_per_read_roc ` command limits included statistics to only those positions found within a specific sequence motif. This is intended to be used in bacterial methylation comparisons between...

The option should look as follows: `--motif-descriptions A:1:A` The `--per-read-statistics-filenames` and `--per-read-control-statistics-filenames` options define the control and sample statistics for this command. You can run any tombo command with the...

I would agree with your analysis. This analysis pipeline has resulted in relatively poor results for 6mA detection. Changing the log likelihood thresholds may help to an extent, but this...

Looking at the precision/recall curve once more, it does appear that a small fraction of sites ~5% do show strong separation from the control sample, while the remainder appear quite...

This is actually another special case that may be effecting your results. Modifying all of a particular base to another base has more global effects on both the signal normalization...