Rna-Seq Analysis Of 454/Roche Data

Question

[raj.gzra]

hi everyone i am having roche/454 transcripts reads in two different condition. i want to check the gene expression. whether it is good to use TopHat and cufflinks for 454/roche data.or suggest me any other software.

[Damian Kao]

How many reads do you have per sample?

Answer 1

[vinay kumar]

I am not sure if TopHat -> Cufflinks is a good choice since 454 does not generate enough throughput as Illumina and SOLiD. You need to estimate the expression using Tag count methods (such as HTSeq followed by DGESeq or DE-Seq or some other simple statistical test (Fisher's exact test??). There have been several papers using EST data to estimate the expression. You can follow their protocols. Try this: http://www.ncbi.nlm.nih.gov/pubmed/20043850

[raj.gzra]

thank you i will try this

Answer 2

[JC]

I will use Blat/BWA-Mem to align the reads, TopHat is not designed to map 454 sequences (long and variable reads, errors, ...), then you can use Cufflinks or HTSeq+EdgeR/DESeq for DEG.

[raj.gzra]

i mapped transcripts reads to already assembled transcriptome with BWA and when samtools flagstat command used, it shows number of input reads is higher than that i have. when i did it with bowtie and then samtools flagstat command used, it shows number of input reads equal to that i have. why samtools flagstat command shows higher number of input reads when reads mapped with BWA???