Daniel:Notebook/GenomeMiner/2013-8-27: Difference between revisions

From ZhangLabWiki
Jump to navigation Jump to search
>Djacobse
No edit summary
>Djacobse
 
(7 intermediate revisions by the same user not shown)
Line 29: Line 29:
'''Replace Matched pairs with Equals'''
'''Replace Matched pairs with Equals'''
  samtools calmd -eS hrcp_sorted_filtered_hl155.sam ../probeseq/Probelist_all.fa > hrcp_sortfilt_eq.sam
  samtools calmd -eS hrcp_sorted_filtered_hl155.sam ../probeseq/Probelist_all.fa > hrcp_sortfilt_eq.sam
So apparently I don't need to use all the indexes.  I just used indexes 10-13, which means I can skip everything but those.  Using all the indexes might screw up some of the statistics, mostly if someones sequence was close enough to have an alignment (a bad one, most likely).
==Phred 64, Shaved Indexing==
'''Bowtie2 with Phred 64 and Fewer Indexes'''
bowtie2 -x hrcp_probesequences -U s_2_1_Indx12.txt,s_2_1_Indx13.txt,s_3_1_Indx11.txt,s_3_1_Indx10.txt,
s_3_1_Indx12.txt,s_2_1_Indx11.txt --phred64 -S hrcp_slimindex_samout_p64.sam
'''Results'''
55964494 reads; of these:
  55964494 (100.00%) were unpaired; of these:
    35040039 (62.61%) aligned 0 times
    19308407 (34.50%) aligned exactly 1 time
    1616048 (2.89%) aligned >1 times
37.39% overall alignment rate
If you look above, you'll see that ~20000 less reads aligned exactly one time and about 1500 less reads aligned greater than one time, meaning about 21500 reads were aligned that shouldn't have been when I was using the full index.  As such, I'm going to continue the pipeline but using the sam files for the fewer indexes.
'''Sam > Bam and Sort Bam'''
samtools view -bS hrcp_slimindex_samout_p64.sam | samtools sort - hrcp_sorted_hl155 &
'''Filter and Convert back to Sam'''
samtools view -h -F 4 -q 8 hrcp_sorted_hl155.bam > hrcp_sorted_filtered_slimindex_hl155.sam &
'''Replace Matched pairs with Equals'''
samtools calmd -eS hrcp_sorted_filtered_slimindex_hl155.sam ../probeseq/Probelist_all.fa > hrcp_sortfilt_slimindex_eq.sam

Latest revision as of 17:27, 28 August 2013

HL155[edit]

Back to Calendar

Running Bowtie2 with Phred 64[edit]

Running bowtie2 again with phred 64 to start the pipeline again. Tried last week but it didn't go so well. I got an error. So I'm running the program again.

Bowtie2 with Phred 64

bowtie2 -x hrcp_probesequences -U  
s_2_1_Indx12.txt,s_2_1_Indx53.txt,s_2_1_Indx54.txt,s_2_1_Indx04.txt,s_2_1_Indx03.txt,s_2_1_Indx05.txt,s_2_1_Indx02.txt,
s_2_1_Indx13.txt,s_2_1_Indx01.txt,s_2_1_Indx55.txt,s_2_1_Indx49.txt,s_3_1_Indx11.txt,s_3_1_Indx10.txt,s_2_1_Indx56.txt,
s_2_1_Indx52.txt,s_3_1_Indx12.txt,s_2_1_Indx11.txt,s_2_1_Indx50.txt --phred64 -S hrcp_fullindex_samout_p64.out

Results:

71613792 reads; of these:
 71613792 (100.00%) were unpaired; of these:
   50674903 (70.76%) aligned 0 times
   19321427 (26.98%) aligned exactly 1 time
   1617462 (2.26%) aligned >1 times
29.24% overall alignment rate

Sam > Bam and Sort Bam

samtools view -bS hrcp_fullindex_samout_p64.sam | samtools sort - hrcp_sorted_hl155 &

Filter and Convert back to Sam

samtools view -h -F 4 -q 8 hrcp_sorted_hl155.bam > hrcp_sorted_filtered_hl155.sam &

Replace Matched pairs with Equals

samtools calmd -eS hrcp_sorted_filtered_hl155.sam ../probeseq/Probelist_all.fa > hrcp_sortfilt_eq.sam

So apparently I don't need to use all the indexes. I just used indexes 10-13, which means I can skip everything but those. Using all the indexes might screw up some of the statistics, mostly if someones sequence was close enough to have an alignment (a bad one, most likely).

Phred 64, Shaved Indexing[edit]

Bowtie2 with Phred 64 and Fewer Indexes

bowtie2 -x hrcp_probesequences -U s_2_1_Indx12.txt,s_2_1_Indx13.txt,s_3_1_Indx11.txt,s_3_1_Indx10.txt,
s_3_1_Indx12.txt,s_2_1_Indx11.txt --phred64 -S hrcp_slimindex_samout_p64.sam

Results

55964494 reads; of these:
 55964494 (100.00%) were unpaired; of these:
   35040039 (62.61%) aligned 0 times
   19308407 (34.50%) aligned exactly 1 time
   1616048 (2.89%) aligned >1 times
37.39% overall alignment rate

If you look above, you'll see that ~20000 less reads aligned exactly one time and about 1500 less reads aligned greater than one time, meaning about 21500 reads were aligned that shouldn't have been when I was using the full index. As such, I'm going to continue the pipeline but using the sam files for the fewer indexes.

Sam > Bam and Sort Bam

samtools view -bS hrcp_slimindex_samout_p64.sam | samtools sort - hrcp_sorted_hl155 &

Filter and Convert back to Sam

samtools view -h -F 4 -q 8 hrcp_sorted_hl155.bam > hrcp_sorted_filtered_slimindex_hl155.sam &

Replace Matched pairs with Equals

samtools calmd -eS hrcp_sorted_filtered_slimindex_hl155.sam ../probeseq/Probelist_all.fa > hrcp_sortfilt_slimindex_eq.sam