I have HiSeq data from mice exposed to two conditions. I would like to answer the following question: "Is there a significant difference in mRNA transcript levels when comparing condition A to condition B?"

To answer this question, do I have to map my HiSeq reads to the reference genome to determine expression values or is it acceptable to use a small set of reference sequences of interest (approx. 50)? These particular sequences are not annotated on the reference genome. What are the downsides to not mapping to the genome?

  • $\begingroup$ how many mice are in your groups? do I understand you correctly that you have 50 mice in your reference group? $\endgroup$ – Vance L Albaugh Jul 7 '16 at 21:46
  • $\begingroup$ Seven in each group of mice. I would like to examine transcript levels for 50 mRNA fasta sequences. Essentially, can I "fish" for only these 50 sequences out of my Illumina Hi-Seq data without mapping to the mouse genome first? Will a lack of mapping to the genome falsely inflate what is identified as one of my sequences of interest? $\endgroup$ – user24933 Jul 7 '16 at 22:24
  • $\begingroup$ Does this help? biostars.org/p/75462 It's a post about aligning sequences using a reference genome without annotating. Or this article about de novo transcriptome assembly: genomebiology.biomedcentral.com/articles/10.1186/… $\endgroup$ – Gaius Augustus Jul 7 '16 at 22:24
  • $\begingroup$ @VanceLAlbaugh Thank you for your thoughtful edit and clarifying questions $\endgroup$ – user24933 Jul 8 '16 at 0:06
  • $\begingroup$ @GaiusAugustus The biostars post is helpful, thank you. Based on that post it seems that what I am proposing is the "quick and dirty way to do basic gene expression experiments", albeit perhaps less excusable given that there is a great deal of information known about my organism versus that of the biostars post. I feel that de nevo assembly will not be necessary for my purposes but thank you for the reference. I will proceed both with and without the genome mapping to compare. $\endgroup$ – user24933 Jul 8 '16 at 0:54

It's going to be hard for you to generate an RPKM or RPM value without % aligning to the genome, or the transcriptome. Your read counts will have no context.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy