Im trying to write my first IRB for a Trio based study. I've done this work in the past and I use a workflow provided by golden helix. MY IRB was sent back because they didn't like the statistical analysis section. I've scrapped what I had and re-wrote it.... is this IRB worthy? Im certainly no biostatistician.
"We will test SNPs for association to susceptibility to severe osseous deformity of the lower extremity by means of a family-based design. Qualifying patients will be designated as cases and respective parents as controls. Homozygous substitutions and heterozygous de novo variants will be analyzed if they are rare (below 0.5% frequency in 1000 genomes). De novo variants will be detected by analysis of whole-exome DNA sequencing data from a family trio study. Typically, these studies generate ~100,000 variations per trio. In an advanced workflow, a combination of functional and sequence quality control measurements will be used to filter the DNA sequence data to obtain a small number of candidate de Novo, rare-recessively inherited, and compound heterozygous mutations. GenomeBrowse and SVS 8 (Golden Helix) will be used to decrease the number of false positive candidate polymorphisms. Variants found in exome data will be filtered to a manageable number based on family structure, protein altering capabilities, and data quality (based on allele depth, read depth, and genotype quality score)."
What do you expect that manageable number to be? This number, along with your target sample size, will heavily influence your statistical power (i.e. the rough likelihood that you'll be able to find your "true positive" result if it is there). They will also be looking for how you are planning to asses the significance of each candidate variant once you have finished filtering the exome data.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics