Hi, i am a post-doc who recently turned to bioinofrmatics. My institute has a nice facility, but my lab is very inexperienced so I could use some external help!
Basically I will assemble transcriptomes from several individuals of my species de novo (no reference available). I wish to compare each individuals transcriptome and isolate polymorphic genes.
I have found several examples of this with people looking for SNPs. But when I say polymorphism, I mean hyperpolymorphism (upto 30% difference for example, but potentially interspersed throughout specific regions of the gene). So definitely the same gene, but these genes should be significantly polymorphic from that of another individual.
Most approaches seem to pool reads from all available samples to form the reference transcriptome, then map each individual sample to the reference. I can understand that this would work for SNPs but I am worried if we have genes that are hyper-polymorphic, they may appear as separate transcripts altogether in any reference transcriptome and thus the reads would not stack in regions of genuine polymorphism.
I have PE 100bp reads with illumina, I have not got the data back, but coverage should not be an issue. Assembly is done with Trinity. If I have a hyper-polymorphic gene, but it is highly conserved for example at one end, will my resultant reference transcriptome be sensitive enough to only produce a single contig for this hyperpolymorphic gene and allow my reads to be stacked accurately?
Or is there a better way to go about it? Thanks for any help to an NGS novice!!
SJLH
Basically I will assemble transcriptomes from several individuals of my species de novo (no reference available). I wish to compare each individuals transcriptome and isolate polymorphic genes.
I have found several examples of this with people looking for SNPs. But when I say polymorphism, I mean hyperpolymorphism (upto 30% difference for example, but potentially interspersed throughout specific regions of the gene). So definitely the same gene, but these genes should be significantly polymorphic from that of another individual.
Most approaches seem to pool reads from all available samples to form the reference transcriptome, then map each individual sample to the reference. I can understand that this would work for SNPs but I am worried if we have genes that are hyper-polymorphic, they may appear as separate transcripts altogether in any reference transcriptome and thus the reads would not stack in regions of genuine polymorphism.
I have PE 100bp reads with illumina, I have not got the data back, but coverage should not be an issue. Assembly is done with Trinity. If I have a hyper-polymorphic gene, but it is highly conserved for example at one end, will my resultant reference transcriptome be sensitive enough to only produce a single contig for this hyperpolymorphic gene and allow my reads to be stacked accurately?
Or is there a better way to go about it? Thanks for any help to an NGS novice!!
SJLH