problem aligning SOLiD reads with Bowtie

[pgalante]

Hi all,

I'm using bowtie to align SOLiD reads against human genome, but the result looks non-sense for me. If you align (bowtie) the read listed below against the chr11 (hg19), for example, bowtie will decode the read (color code to base code) in very different nucleotide (see example of alignment listed below). It looks to me that bowtie is not doing the correct thing that is align the read in color space and then decode to base space. Anyone had the same problem? Suggestions to solve the problem?

===> read <===
@1_211_1749_5_SR1_F3
T01001001000000000000000000000000000000000000000000
+
(#$##$##$###'#$##'##)&+''#(,)((#,,+##,,)+2),,,#,'+

===> result of alignment <========
1_211_1749_5_SR1_F3 - chr11 97249435 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACCACCC 12..774:<347.%-67.*/043*)-10.+%))%&&))%%&&%!!%&& 0
1_211_1749_5_SR1_F3 + chr11 41639884 TTTGGCAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA "&%&"!&&%%))&&%))%+.01-)*340/*.76-%.743<:477..21 0
1_211_1749_5_SR1_F3 + chr11 33273211 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA &&!"&!"&%%))&&%))%+.01-)*340/*.76-%.743<:477..21 10
1_211_1749_5_SR1_F3 + chr11 70107678 TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT &&!"&!"&%%))&&%))%+.01-)*340/*.76-%.743<:477..21 10

[adamdeluca]

Could this is caused by single color errors messing with the phase of the colorspace translation?

I would try:
http://solidsoftwaretools.com/gf/project/saet/
to correct the one color errors before aligning with bowtie

[nilshomer]

Hi all,

I'm using bowtie to align SOLiD reads against human genome, but the result looks non-sense for me. If you align (bowtie) the read listed below against the chr11 (hg19), for example, bowtie will decode the read (color code to base code) in very different nucleotide (see example of alignment listed below). It looks to me that bowtie is not doing the correct thing that is align the read in color space and then decode to base space. Anyone had the same problem? Suggestions to solve the problem?

===> read <===
@1_211_1749_5_SR1_F3
T01001001000000000000000000000000000000000000000000
+
(#$##$##$###'#$##'##)&+''#(,)((#,,+##,,)+2),,,#,'+

===> result of alignment <========
1_211_1749_5_SR1_F3 - chr11 97249435 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACCACCC 12..774:<347.%-67.*/043*)-10.+%))%&&))%%&&%!!%&& 0
1_211_1749_5_SR1_F3 + chr11 41639884 TTTGGCAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA "&%&"!&&%%))&&%))%+.01-)*340/*.76-%.743<:477..21 0
1_211_1749_5_SR1_F3 + chr11 33273211 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA &&!"&!"&%%))&&%))%+.01-)*340/*.76-%.743<:477..21 10
1_211_1749_5_SR1_F3 + chr11 70107678 TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT &&!"&!"&%%))&&%))%+.01-)*340/*.76-%.743<:477..21 10

Be more specific about your concern. It looks fine to me, since the color encoded version of those alignments nearly match the origin color sequence. Decoding to base space means that color errors and variants are detected relative the local reference, which is different for each of the alignments above.