Ylaine/2009-8-6
Jump to navigation
Jump to search
Error Classification[edit]
- Write script to tag SNP mismatches from gold standard set as type I or II errors: 'fnErrorType.pl'
../Scripts/fnErrorType.pl fn.1K snp.chr.txt>fn.1K.errors
- Only 87/2910 errors are type 1 at coverage >1 (found 2898 errors on 7/29??)
- Write script to return SNPs above coverage threshold, 'threshold.pl'
../Scripts/threshold.pl fn.1K.errors 8 > fn.8.temp
' | Type 1 | Type 2 | Total |
2x | 87 | 2823 | 2910 |
8x | 23 | 221 | 244 |
10x | 21 | 137 | 158 |
20x | 8 | 27 | 35 |
- Alternatively:
' | Type 1 | Type 2 |
2x | 2.99% | 97.01% |
8x | 9.43% | 90.57% |
10x | 13.29% | 86.71% |
20x | 22.86% | 77.14% |
False Positives[edit]
- Effect of probe efficiency
/Users/kunzhang/WorkSpace/Exome/Solexa/NA12878/Exome_Probe_efficiency.txt
- Use grep to categorize efficiency into 6 bins:
~/NA12878/Probes
Bin # | Efficiency | Number |
1 | [1,20] | 138835 |
2 | [21,40] | 27059 |
3 | [41,60] | 9714 |
4 | [61,80] | 4027 |
5 | [81,100] | 1813 |
6 | >100 | 1247 |
- Writing 'calculateSNPeff.pl'
- Assume ranges are inclusive of region strictly between target arms