Probably a naive question. I am inexperienced.

I am interested in identifying potential CRISP (Cysteine-rich secretory proteins) in a certain tissue transcriptome (ca. 20k sequences in fasta). I have detected signalP and estimated % of cysteine in sequences. However CRISP rely on a certain pattern, and I am not aware of batch-detected algorithms.

Please would anyone know the best/simplest way of detecting CRISP candidates?


I have been asked for the pattern. I summarise what the literature says:

  • Several of them have a CAP domain;

  • There should be a region called 'cysteine-rich domain' (CRD) upon the carboxyl terminal half of the protein containing 10 of roughly 16 conserved cysteines.

I am particularly interested in venom-derived CRISPs, but I am not sure whether they cluster in a certain CRISP family.

References: Roberts et al. (2007) Structure and function of epididymal protein cysteine‐rich secretory protein‐1. Yamazaki & Morita (2004) Structure and function of snake venom cysteine-rich secretory proteins.

  • 2
    $\begingroup$ Could you edit your question and tell us what this pattern is? Is it a specific sequence pattern? If you can define the pattern, we could be able to figure something out even if there aren't any existing dedicated tools for this. $\endgroup$
    – terdon
    Jul 2, 2018 at 10:02
  • $\begingroup$ @terdon I have read some references and hopefully the pattern described in my edit update comprises what I am looking for. Thanks $\endgroup$
    – Scientist
    Jul 3, 2018 at 17:35

1 Answer 1


Is your goal to detect the actual proteins from the sequences? If you'll settle for detection of transcripts from the genes encoding these proteins as opposed to detection of the proteins themselves, here is an approach.

Aligning your transcriptome to a reference genome (probably with STAR) will be the first step. You have lots of short reads in your FASTA, but in their current state they are no more informative than a box of puzzle pieces. Aligning your reads to the reference will allow you to extrapolate which genes that you short reads correspond with. From here you can quantify the features (genes) in your transcriptome.

Lucky for you, there are only 3 genes in the cysteine rich secretory protein family, CRISP1, CRISP2, and CRISP3. Alignment with a reference (and annotations) and quantification of genes that your reads align to will give you all you need. You'll just have to find the three genes in your final counts table.

I recommend STAR as an aligner with the --quantMode GeneCounts argument. It's very quick if you've got the resources and is geared towards RNA-seq reads.

  • 1
    $\begingroup$ My fasta is the result of de novo assembly, plus we have a reference genome from a proximate species for comparison. Problem is, these are not intensively-annotated organisms (e.g. not Drosophila nor Mus musculus) thus I am trying to fish out as many annotations of novel transcripts as possible. $\endgroup$
    – Scientist
    Jul 3, 2018 at 17:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.