t**k
发帖数: 16 | 1 新人求教,现在手里有上千样本的病人和正常人的genotypes数据,用什么软件,或者有什
么办法能够分析得到更加准确的hyplotypes,并且进一步找到可能出问题的DNA片段所在
位置? | e*****t
发帖数: 642 | | b****r
发帖数: 17995 | 3 免费的haplopview就可以估计haplotype，用的人也非常多。据我所知好像也谈不上什
么“更准确的haplotype”，大家没有几个人真正去大规模验证吧，都是refer出来的，
方法上各有优势，都有不少人用。
或者打算投啥杂志，看看主编的文章一般用啥软件 | n********t
发帖数: 1079 | 4 Haploview，不过数据量不能太大，这个软件的memory footprint大得惊人 | t**k
发帖数: 16 | 5 haploview虽然能给出一些risk haplotype, 但是老板让我作功能相关的研究，验证这
些结果，非常挠头，不知道从哪里下手，有人跟我说，针对p值高的分析一些dna 蛋白
结合，EMSA, 但是我感觉很多SNPs都很难说，有点瞎猫碰死耗子的感觉。
老板一心想深入下去，而且还在作进一步的sequencing，并且给我很多比hapmap和
1000genom project 里面更全的genotypes data。 弄得我都不知道怎么利用这些东西
了。 | s*****0
发帖数: 357 | 6 This is a really good opportunity to get exposure to the knowledge of the
association study. If you are interested in an industry job, this project
will definitely help.
You'd better formulate your question more clearly. What is your plan? Why
going directly to haplotypes without conducting single site tests? How large
and how dense of the coverage you want your haplotype to be? How would you
plan to use the statistical inference to guide your functional study?
Different software (PHASE, PLINK, HAPLOVIEW etc.) or haplotyping algorithms
(EM or Bayesian) have their pros and cons. You need to understand the
limitations before you attempt to apply them. Some are baised toward common
haplotypes but others might favor rare haplotypes. What is your desired
cutoff for the minimum frequencies of single SNP or haplotypes? I believe
that you should have a clear hypothesis and project scope before you attempt
any analyses. Otherwise, the study scale will be unbounded, and the
intended scientific exploration will end up as data dredging.
【在 t**k 的大作中提到】
: 新人求教,现在手里有上千样本的病人和正常人的genotypes数据,用什么软件,或者有什
: 么办法能够分析得到更加准确的hyplotypes,并且进一步找到可能出问题的DNA片段所在
: 位置?
| t**k
发帖数: 16 | |
|
