Hello!

It depends on how refined you need the analysis to be.

1) You can do it on servers like:

PhaGAA (http://phage.xialab.info/home)

PHP (http://www.computationalbiology.cn/phageHostPredictor/home.html).

PhaBOX (https://phage.ee.cityu.edu.hk/)

HostPhinder (https://cge.food.dtu.dk/services/HostPhinder/)

2) If you have a good bioinformatics resource, you can use tools like:

CHERRY (https://github.com/KennthShang/CHERRY)

iPHoP (https://bitbucket.org/srouxjgi/iphop)

PHERI (https://github.com/andynet/pheri)

DeepHost (https://github.com/deepomicslab/DeepHost)

I share with you this interesting literature: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8778386/

These tools have another purpose but could be useful

PhageAI (https://phage.ai/)

VICTOR (https://ggdc.dsmz.de/victor.php)

Greetings

I don't think you can predict host range using bioinformatics tools There are so many subtleties that impact host range, both in terms of gene expression, chaperons that influence folding of proteins, and most importantly the interactions with the host receptor for phage binding and injection. We don't yet know enough to predict except in a very few well studies examples.

Michael J. Benedik

Hello,

Thank you for your valuable contributions. There are some points in your answer that have confused me. It is understandable that there are numerous factors that can affect host range. Many studies in the literature begin by examining phage receptor binding proteins when assessing host range. The rationale behind this is that the success of the remaining processes of the phage depends on the recognition of the host by these RBPs. For example, if a phage cannot recognize its host, how will it inject its genome? There is a known scenario, too where the phage may recognize the host, but the genome may not be successful enough to infect the bacteria. However, in my case, I see lytic zones on the petri dish, which I believe indicates a successful infection process. Therefore, in the case of my phage, I can at least predict that a successful host recognition and lytic process has occurred.

I have a phage that infects different species. Since it was able to carry out a successful infection, can we conclude that it has performed the RBP recognition process? We also know that the membrane and cell wall compositions of Gram positive and Gram negative bacteria are different from each other. Based on the literature, some phages are said to provide their host range with a single RBP, while others may have more than one tail protein. The concept of a single RBP for a phage that recognizes different species within the same genus makes sense because there are more conserved target receptor portions among Gram-negative bacteria. Therefore, it seems likely that the phage would also recognize them with minor changes in the RBP. However, wouldn't it be too difficult for a phage to infect both Gram-positive and Gram-negative bacteria? I have observed two RBP proteins in the phage I am studying. I speculated that these RBPs might be effective in recognizing heterogeneous hosts, so I expected to find a connection with the phage infecting a Gram-positive host, at least through one of the RBPs. (My host is Gram-negative)

Additionally, I expected to find a connection between lytic enzymes and a successful infection scenario. The difference in the wall compositions of Gram-negative and Gram-positive bacteria causes endolysins to exhibit different domain properties. The phage I am studying has a similar level of endolysin as seen in phages infecting Gram-negative hosts. Although I haven't focused on holin, I am interested in understanding how the endolysin affects the membrane structure of S. aureus. We know that the folding of proteins is determined by sequence information and that chaperones have no effect on protein folding in bacteria. Protein folding, in turn, affects protein function. Therefore, while investigating the relationship between functional recognition and lysis proteins, shouldn't the sequence information also provide insights? In my scenario, shouldn't I be able to establish some connections at the genome level without solely examining the molecular mechanisms, considering that the phage already shows a genus-level host range in vitro?

I hope this clarifies my questions and concerns. Thank you for your assistance.

Hi Berna - you raise some good points and let me be clear that my answers are only speculation. I am quite surprised that you found a phage capable of infecting both Pseudomonas and Staph, I've never heard of that.

If you are asking whether you can tell if a phage's host is likely to be G+ or G-, then it is possible you can distinguish that by bioinformatics for some of the reasons you mention. The lysis cassette is likely to be a bit different (no need for holins in G+ for example. But I don't think bioinformatics could distinguish whether the phage infects Staph vs Bacillus or E. coli vs Pseudomonas. In my earlier answer I was more addressing the latter, just because you find an RPB that is somewhat similar to an E. coli phage doesn't mean you can at all predict what receptor it binds to and how tight that specificity is among other G- bacteria.

In fact some phages (Mu for example) has 2 different receptors for different hosts and switches between them. Similarly single amino acid change in some RPBs can alter (extend or limit) the host range of some phages. Some phages don't even bind to protein receptors, some bind to LPS for example. So there is a wide range of binding targets for the phages.

Just getting the phage genome in is not sufficient, there are some nice studies where you can introduce the E. coli lamB gene into a variety of hosts and phage lambda will bind and inject (as shown by transduction of broad host range plasmids) but will not form a plaque. So this is some regulatory/gene expression issue.

So of course you can learn things from the genome but you need to be cautious. For example in the case of your phage it seems like if you only looked at the genome you might predict it was a G- dependent phage but would likely not predict it also infected G+. But your experiments showed that it did.

I hope this helps a little bit, but it sounds like you have an interesting phage to study.