RNA seq with 2 biological replicates

Question

[mursyidahkhan]

Hello everyone. I'm working on the transcriptome of parasitic plant, Rafflesia sp. My question here is I have only 2 biological replicates due to the limitation to get the samples from wild. I need your help to share with me some journal links that were published with 2 or no biological replicates, so that I can have a better view on this matter. Thanks in advance.

[grant.hovhannisyan]

what is the goal of your study? Biological replicates are critical for dif. expression analysis. For other let's say more descriptive goals biological replicates are of course desirable but are less important than in dif. expression.

[mursyidahkhan]

The idea of the project is to look into multi-transcriptome (the sample that I'm working on is a haustorium sample,which hypothetically has both host and parasite tissue) data and yes, I'm not focusing my study more into DEG since DEG would be more relevant if there's 12 and above bio reps. I would go more towards gene discovery on the host-parasite interaction pathway.

[grant.hovhannisyan]

Another question, are there reference genomes available for both species (host and parasite)? If yes, then your life is easier - sequence your samples with high sequencing depth (for a proper transcript reconstruction for a human, you would typically need around 80-100 mln reads, but I don't know if there are any standards for plants) and relatively long reads (let's say 2x100 bp or more). Also take into account that if your samples contain data from two different species, reads from one species potentially can map to the other species, and vice versa (this is called crossmapping). If you have reference genomes, you can test if this happens using Crossmapper software that we developed in our lab (it is based on simulation so you can run it before doing any sequencing to adjust your sequencing configuration). When everything is fine, you can reconstruct the transcriptomes using HISAT/STRINGTIE pipeline for example. If you don't have reference genomes, then basically you do the same sequencing configuration (high depth + long reads) and then good luck with de-novo transcript reconstruction (can be done by Trinity for example).

[ATpoint]

That is simply a question of effort to go through PubMed and find appropriate papers.

so that I can have a better view on this matter

What does that mean?

Answer 1

[mark.ziemann]

2 replicates is not very reliable. You will have a high rate of false positives and false negatives. That is just a fact of life.

Now with that out of the way, you could read the following:

• EdgeR paper, which talks about small number of replicates https://academic.oup.com/bioinformatics/article/26/1/139/182458
• This evaluiation of replicate number https://rnajournal.cshlp.org/content/22/6/839.long

[mursyidahkhan]

Yes, I understand the situation but this project is more into gene discovery than looking into the DEG (which I know is not reliable for me to conduct). I have gone through some of the journals on the bio reps before and thanks for links too.

[ATpoint]

Conduct your analysis as planned and try to validate the key findings using independent experimental techniques, be it Northern blot, qPCR, long read sequencing etc. This is imho the best way of dealing with uncertainty (and good scientific practice).

[mark.ziemann]

Here is another paper that might be interesting where they discuss RNA-seq without replicates GFOLD: a generalized fold change for ranking differentially expressed genes from RNA-seq data.

Answer 2

[ashish]

I hope you know the importance of replicates in transcriptome studies.

Having said that you need to search for transcriptome studies on plants where sample collection is difficult. You should search the web for such plants. Don't expect people to do the search for you. Litchi is one such plant. Flower collection is very difficult. I have read some studies on litchi ovule with no replicates. There are other plants similar in characteristics to Rafflesia and are also found in the same region. There are published transcriptomic studies with 2 replicates on them.