Daniel:Notebook/GenomeMiner/2013-9-19
Mock HL155 (Started 9/9/2013)
Single Error per Read
Goal of this test is to see if bowtie's alignment results are dependent on the number of errors per read. So far the errors have been completely random, with each base potentially getting a substitution, insertion, or deletion, or even several errors. This new iteration rolls to see if an error will occur at all, and based on its roll also determines which of the three error types it will give.
Workflow
1. MockHL155_Master.m, Switch 6 2. scp v4s1mockseq_1errperread_mimic.fq djacobse@132.239.135.41:/media/LTS_15T/DEJ_LTS/SeqStore/130628_HL155/mockseq/error1per/ 3. hl155bash.sh 4. perl imp_count_mismatch.plx (Matt's error counting script)
MATLAB Error Count:
1494000 substitutions (1.01 pct) 384556 insertions (0.26 pct) 73592 deletions (0.05 pct)
Alignment Results
2959000 reads; of these: 2959000 (100.00%) were unpaired; of these: 25319 (0.86%) aligned 0 times 2928519 (98.97%) aligned exactly 1 time 5162 (0.17%) aligned >1 times 99.14% overall alignment rate
Error Counting Results
Error Rate: 1.298% Error Rate of Insertions: 0.241% Error Rate of Deletions: 2.45e-04% Error Rate of Substitutions: 1.057%
Use Real Bases For Deletions
Previous iterations have used the read length sequences as fasta files. Therefore, deletions add on a random base at the end, since we didn't have data for what the base actually should be. To verify this approach (maybe correct it), I have added a parameter in "gen_fastq_error_range.m" that can use the full length sequence. For this approach, the file only prints bases 1:50 in the output file. Since the reference sequence is longer than 50 bp, if a base is deleted the next base in the reference sequence is automatically added into the output.
Workflow
1. MockHL155_Master.m, Switch 5 (param reflength,full) 2.