genomics question 1. there are huge numbers of SNPs thats tricky 2. there are 23 or so chromosomes 3. roughly, chromosomes can change by mutation or by recombination 4. recombination is relatively infrequent for adjacent sequences 5. mutation is relatively infrequent
-
Show this thread
-
so A. seems like within a population there will probably a relatively (compared to SNPs especially) small number of consistently-appearing longer segments across generations, especially if some loci are quite a lot more likely to be the site of crossing over than others?
1 reply 0 retweets 5 likesShow this thread -
B. if you can identify these groups you might have a better time using them as features (vs SNPs) especially given eg role of introns in the regulation of expression?
2 replies 0 retweets 2 likesShow this thread -
C. Seems like you could deduce the traits of (successful) founders of populations from these traits a la linguistic archaeology?
4 replies 0 retweets 6 likesShow this thread -
Replying to @eigenrobot
Yeah, the term you're looking for here is haplotype: https://www.genome.gov/genetics-glossary/haplotype … Some of what you suggest is possible, but tricky. Key issue is that building haplotypes is pretty hard with current sequencing tech.
2 replies 0 retweets 13 likes -
Replying to @arguablywrong @eigenrobot
The old/cheap style sequencing just picks up known (i.e., common and easy to track) point mutations that can be really far from each other. If someone has A/T at SNP1 and G/C at SNP2, are the haplotypes A-G & T-C or A-C and T-G?
1 reply 0 retweets 3 likes -
Replying to @arguablywrong @eigenrobot
This is what 23andme does, and you basically can't phase a genome (i.e. generate haplotypes) this way, unless you have multiple family members.
2 replies 0 retweets 5 likes -
Replying to @arguablywrong @eigenrobot
To expand on the family member thing - suppose you have the SNP data on mom, dad, and multiple of their kids. For best results, get a mom and dad that are wildly genetically different (i.e., unlikely to share large haplotypes.)
1 reply 0 retweets 2 likes -
Replying to @arguablywrong @eigenrobot
Take a kid. They've got exactly one of their dad's two haplotypes and exactly one of their mom's, but recombined so that the two haplotypes are scrambled. So you can't phase the kid's haplotype. But you can use the kid's data to phase the *parents*.
2 replies 0 retweets 2 likes -
Replying to @arguablywrong @eigenrobot
Example: Mom - 1:A/T 2: G/C 3: T/T 4: C/C Dad - 1: C/C 2: A/A 3: T/A 4: G/C Kid - 1: C/A 2: A/C 3: T/T 4:C/G You can then deduce that the mom's haplotypes are A-C-T-C and T-G-T-C and the dad's are C-A-T-G and C-A-A-C. Repeat with more kids for more reliable results.
2 replies 0 retweets 1 like
Yeah ok I'm seeing it
Tricky tricky problem
Very cool 
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.