nanoporetech
Provide New Feature For Pileup Command：Filter by base_qual
In our daily usage, we may filter out low-confidence data using base_qual and then examine the overall 5mc modification levels through pileup. Therefore, we hope to add an option in the pileup command to provide the functionality of filtering using base_qual.
Hello @WeiHea,
In Nanopore data often times the base_qual is correlated with modification probability. So filtering on base_qual could bias your modification estimates. Do you find this to be the case in your analysis? You could (and probably should) filter the modBAM based on quality (qs:f: tag) or other read-level quality metrics.
Hello @WeiHea,
In Nanopore data often times the base_qual is correlated with modification probability. So filtering on base_qual could bias your modification estimates. Do you find this to be the case in your analysis? You could (and probably should) filter the modBAM based on quality (
qs:f:tag) or other read-level quality metrics.
Are you suggesting that the electrical signals from modifications might lead to lower base_qual? Actually, in our data, after filtering out bases with base_qual ≤ 30, the 5mC modification results became more consistent with our expectations.
Hello @WeiHea,
In Nanopore data often times the base_qual is correlated with modification probability. So filtering on base_qual could bias your modification estimates. Do you find this to be the case in your analysis? You could (and probably should) filter the modBAM based on quality (
qs:f:tag) or other read-level quality metrics.
Additionally, if data with lower base_qual could still represent high-quality modification calls, what alternative approaches would you recommend besides using adaptive thresholding to filter out low-quality data?