Entering edit mode
5.4 years ago
lffu_0032
▴
90
hi, everyone, i always have a question about the detailed algorithms used in gatk re-alignment step. i looked through the GATK documents but did not find useful information. the only thing useful i found is in the GATK wrokshop, https://drive.google.com/drive/folders/0BwTg3aXzGxEDN0pTMzV2aGFlQ2M, Realignment around indels in 8th ppt：1）Finds minimum alternate consensus sequence(s) that best fits reads；2）score for alternate consensus uses sum of quality scores of mismatching bases; 3)If best alternate consensus is sufficiently beher than the original alignments -> accept proposed realignment。can someone provide more detailed info about re-align in GATK. thanks a lot.
Which step or steps in there do you find confusing? Have you read the source code?
at the first step that i showed above: 1）Finds minimum alternate consensus sequence(s) that best fits reads. i do not know the exact meaning of this sentence. i am going to read the gatk-dev team's nature genetics paper titled "A framework for variation discovery and genotyping using next-generation DNA sequencing data" for more information.
sorry, i am not familiar with the java language, so i did not read the source code. do you have any suggestions? thanks a lot.
I hope you're aware that the tool itself no longer exists. The general idea with realigning around targets is to do construct a de Bruijn graph surrounding an indel. Then the possible paths through that graph are determined and reads compatible with each are counted. Paths can then be eliminated according to how many reads are compatible with them. I can't say for certain that's exactly how GATK implemeted its old IndelRealigner, but that's generally how it works.
I don't why you didn't find the following GATK indel realignment(local realignment) related documentation.
RealignerTargetCreator
IndelRealigner
of course i had looked through this two pages carefully, i mean the detail information of re-align. this two pages just give some basic principles of re-align and show some examples about how to use it, plus the detailed meanings of each parameter. however, i really can not find more explanations about the detailed steps that re-algin involves.