FelixKrueger
commented
1 year ago Hi @ifalconi20
Ahh, you've got Nanopore data. May I ask if the data was bisulfite converted at all, or using a similar method such as EM-seq? Please note that Bismark doesn't just do a methylation analysis of sequencing data, but it is explicitly designed to work with bisulfite-seq data.
Bismark supports the use of minimap2 to align Nanopore reads, I am sure this would be a much better option than tweaking the options for Bowtie2 alignments, see here: https://github.com/FelixKrueger/Bismark/releases/tag/0.24.0
As an example, for a 200bp long read with a --score_min L,0,-21 would allow a penalty score of -4200, or 700 mismatches (!) and or indels, which seems to be somewhat overly lenient. :)
ifalconi20
I have the sequencing data in Nanopore of a small amplicon (230 bp) of the HLA-C gene. One of the challenges I face is the low quality of the sequences obtained, as well as the fragmentation of the sequences (most have the size of the amplicon, but there are others that are smaller, some up to 100 bp and some larger, up to 700-1000 bp). This is because I do not have a specific kit for short sequences, so I had to use a kit designed for long sequences, which caused fragmentation of my sequences due to the action of the transposase.
My goal is to get the most out of the reads obtained, so I performed three sequences using the same method to increase coverage. First, I used the Porechop tool to remove barcodes designed for Nanopore and possible chimeras. Then, I employed another sequence trimming tool called Filtlong to maintain a length range between 44 and 230 bp in order to optimize the mapping process.
Although I am still unsure whether the mapping should be directional or non-directional, I initially performed the mapping in a directional manner. However, I experienced low mapping efficiency, ranging from 7% to 15% in all my samples. After researching in the official guide and consulting some discussions in specialized forums, I decided to use the following command to improve the results:
bismark --score_min L,0,-10 --phred33 -N 1 -L 10 --genome (Referencia) (archivo)
In the hope of increasing mapping efficiency, I adjusted the parameters of the Bismark command. I set a minimum score value (score_min) at L,0,-21, used a phred33 sequence quality format, and allowed up to 1 error in alignment (-N 1). In addition, I set a minimum alignment length threshold at 10 (-L 10). Finally, I indicated the reference genome that I used in the mapping process. With the adjustments made to the Bismark command, I managed to improve the mapping efficiency to a range between 48% and 62%. However, I would like to confirm if the implementation I have carried out is adequate or if another approach or strategy for data handling is needed.
I have also faced problems where some sequences try to align to the end instead of the start of the sequence. Despite adding "NN" to both the beginning and end of the reference sequence, it has not completely solved the problem. It is important to note that this situation occurs in a relatively low number of entries, around 40 or 60 compared to the total number of reads, so the difference is not significant.