Bismark release v0.9.0
We have just released a new version of Bismark (v0.9.0) which most notably introduces a new HTML report to produce a visual summary of the alignment, deduplication and M-bias statistics of a BS-Seq experiment (requires Javascript). Here are examples for a standard paired-end BS-Seq or a single-end PBAT report. It also adds a new Unknown cytosine context for Bowtie 2 alignments for Cs that are very close to Ns or insertions in the reference sequence. Here are all the changes in more detail:
• Bismark: Implemented the new methylation call symbols 'U' and 'u' for methylated or unmethylated cytosines in Unknown sequence context, respectively. If the sequence context bases contain an N, e.g. CN or CHN, the context cannot be determined accurately (previously, these cases were assumed to be in CHH context). These situations may arise whenever the reference sequence contains Ns, or when insertions in the read occur close to a cytosine position (bases inserted into the read have no direct equivalent in the reference sequence and were assumed to be Ns for the methylation call). In practical terms, the 'U/u' methylation calls will only occur for Bowtie 2 alignments because Bowtie 1 does not support gapped alignments or read alignments if the reference contains any N's. The Bismark report will now also include the 'U/u' statistics, such as count and % methylation, however only if run in Bowtie 2 mode.
• bismark2report: this new module generates a graphical interactive HTML report of the Bismark alignment, deduplication, splitting and M-bias statistics for convenient visualisation of what is going on. Since several different modules of Bismark may be included into this report that may or may not have been run, bismark2report requires the user to specify the relevant reports as input files. Many thanks to Phil Ewels for the conceptual design and his help with this report.
• Bismark: Fixed a bug affecting the generation of the alignment overview pie chart which occurred for PBAT libraries only
• Methylation Extractor: Added handling of the newly introduced methylation call U/u for cytosines in Unknown sequence context (CN or CHN). These methylation calls are simply ignored in the extraction process to not cause too much confusion for downstream analysis
• bismark2bedGraph: Added a check to see whether input files start with CpG_* or not. If they don't, please include the option '--CX' when running bismark2bedGraph as a stand-alone tool
The new bismark2report module appears to be working well with the latest version Bismark, but I didn’t have the time to investigate how far backward-compatible it is. Also, error handling or documentation might need some further updating. However, as it is the summer holiday period I won’t be able to fix any potential bugs immediately, but comments/criticism or suggestions are very welcome and I will deal with them upon my return.
Bismark is available for download from https://www.bioinformatics.babraham....jects/bismark/.
We have just released a new version of Bismark (v0.9.0) which most notably introduces a new HTML report to produce a visual summary of the alignment, deduplication and M-bias statistics of a BS-Seq experiment (requires Javascript). Here are examples for a standard paired-end BS-Seq or a single-end PBAT report. It also adds a new Unknown cytosine context for Bowtie 2 alignments for Cs that are very close to Ns or insertions in the reference sequence. Here are all the changes in more detail:
• Bismark: Implemented the new methylation call symbols 'U' and 'u' for methylated or unmethylated cytosines in Unknown sequence context, respectively. If the sequence context bases contain an N, e.g. CN or CHN, the context cannot be determined accurately (previously, these cases were assumed to be in CHH context). These situations may arise whenever the reference sequence contains Ns, or when insertions in the read occur close to a cytosine position (bases inserted into the read have no direct equivalent in the reference sequence and were assumed to be Ns for the methylation call). In practical terms, the 'U/u' methylation calls will only occur for Bowtie 2 alignments because Bowtie 1 does not support gapped alignments or read alignments if the reference contains any N's. The Bismark report will now also include the 'U/u' statistics, such as count and % methylation, however only if run in Bowtie 2 mode.
• bismark2report: this new module generates a graphical interactive HTML report of the Bismark alignment, deduplication, splitting and M-bias statistics for convenient visualisation of what is going on. Since several different modules of Bismark may be included into this report that may or may not have been run, bismark2report requires the user to specify the relevant reports as input files. Many thanks to Phil Ewels for the conceptual design and his help with this report.
• Bismark: Fixed a bug affecting the generation of the alignment overview pie chart which occurred for PBAT libraries only
• Methylation Extractor: Added handling of the newly introduced methylation call U/u for cytosines in Unknown sequence context (CN or CHN). These methylation calls are simply ignored in the extraction process to not cause too much confusion for downstream analysis
• bismark2bedGraph: Added a check to see whether input files start with CpG_* or not. If they don't, please include the option '--CX' when running bismark2bedGraph as a stand-alone tool
The new bismark2report module appears to be working well with the latest version Bismark, but I didn’t have the time to investigate how far backward-compatible it is. Also, error handling or documentation might need some further updating. However, as it is the summer holiday period I won’t be able to fix any potential bugs immediately, but comments/criticism or suggestions are very welcome and I will deal with them upon my return.
Bismark is available for download from https://www.bioinformatics.babraham....jects/bismark/.
Comment