Bowtie-Build, Which File To Use Contigs.Fa Or Transcript.Fa

Question

[nbvasani]

Hi Community users,

I have single-end reads fastq file. I have generated denovo assembly using velvet/oases. Now I need to check how good is my assembly. So I am using aligner, bowtie. But I am confused which fasta file to use for creating bowtie index i.e. shall I used contigs.fa from velveg or transcript.fa file from oases.

Do you guys prefer --best parameter option while running bowtie?

I would really appreciate your feedback and reponse.

Thanks,
Naresh

Answer 1

[Istvan Albert]

Use both and compare the results.

Plus use bwa-mem instead of bowtie, that way your reads will align across junctions even for the contigs.fa file.

[nbvasani]

Thanks Albert. I will try that.

[nbvasani]

Hi Albert, I tried bowtie with contigs.fa file, but I got only 6% alignment. Now I am trying to run with transcripts.fa. Any suggestion why alignment rate is so low. Fastq file used for alignment was trimmed to get better quality.

Thanks in advance. Naresh

[Istvan Albert]

that's why I said use bwa-mem

[nbvasani]

ohh ok. Thanks.

[nbvasani]

Hi Albert,

As per your advice I used bwa-mem for alignment it work out very well. It think I got 79% align. Can you please check my flagstat output. If I am interpreting result in correct manner.

If it is 79% align can you please help me how can I improve alignment above 90% by adding more parameter.

I used following cmd:

##bwa index contigs.fa
##bwa mem contigs.fa CombineIonXpressRNA_009_NareshPool_Chip1_2_WT1_fastx_trimmer_from_quality_trimmer_file.fastq > aln-se.sam
##samtools view -bS aln-se.sam > aligned_reads.bam
##samtools flagstat aligned_reads.bam
34569162 + 0 in total (QC-passed reads + QC-failed reads)
0 + 0 duplicates
27555361 + 0 mapped (79.71%:-nan%)
0 + 0 paired in sequencing
0 + 0 read1
0 + 0 read2
0 + 0 properly paired (-nan%:-nan%)
0 + 0 with itself and mate mapped
0 + 0 singletons (-nan%:-nan%)
0 + 0 with mate mapped to a different chr
0 + 0 with mate mapped to a different chr (mapQ>=5)

Appreciate your help.

Thanks,
Naresh

[Istvan Albert]

there is no reason to expect a 90% alignment rate, some or many of your reads may not have ended up in the assembly

[nbvasani]

Thanks Albert! That mean I understand correctly, 79% was aligned from flagstat output?