What features are available right now?
The basic idea is to enter a marker specific about a cancer and to design a tailored peptide vaccine.
It can find peptides with several properties from either a gene and its mutation, or
from a protein that is generated only by tumour, like Tumour Specific Antigen.
The service is designed to be easy to use and has only a few simple parameters:
It infers a seed for the simulated proteolysis mechanism either from a specific protein tumor marker, or uses a gene and its mutation.
It uses NetMHCpan as well as Uniprot and VaxiJen.
If NetMHCpan is unable to create a short list of peptides, it generates its own.
A limitation is that only the genes that are involved in cancer are usable at the moment.
Another limitation is that not all kind of mutations are implemented,
Complex mutations like insertions, frameshift or chromosome fusions are not supported in this version of the tool.
What is warranted exactly?
Indeed it is the responsability of the oncologist to enter valid information.
The basic idea is to enter a marker specific about a cancer and to design a personalized, low cost peptide vaccine.
Caution: Any information about mutations or markers that are present in "normal" human bodies would be harmful.
It would enable a form of personalized medicine, available through most oncologists.
The oncologist would define a treatment for this patient and that cancer.
The oncologist would then be able to track this marker to see how her patient will react to the treatment.
There is no warranty that a result will be found, simply sometimes there will be no result at all,
especially when there are hard constraints like a specific HLA type.
It is also the responsability of the oncologist to make sure the proposed peptides are not dangerous.
Clearly we have a lot of work to do in this area, please see the roadmap.
Why is it so slow?
First because several online databases are queried sequentially: Uniprot, NetMHCpan, VaxiJen
But also because of the cheap hosting at the moment.
We plan to be less reliant on those databases in the future.
Do you have a roadmap?
Roadmap: Our peptide designs aim at :
...Do not induce anaphylaxis (no B cell epitopes)
...Do not induce autoimmune pathology
...If possible beeing promiscuous (bind to more than one MHC allele)
...Being both CD4 and CD8 epitopes (for efficiency)
...Looking like a Virus Like Particle
...Being oligomeric (looking roughly like a retroviral capsid)
...Having amphilicity (as most viruses)
...Having post-transcriptional modifications fully specified (to standardize production)
...Tailored for intravenous injection
...With references to usage of this peptide in the medical litterature
How to contribute?
The most helpful way is probably to declare issues on Github
It will make it possible to track all bugs and feature request.
Another way is to raise awareness about this effort, particularly in the "off label" usage by oncologists.
Many people are suffering and often dying because they are unable to receive personalized and/or costly cancer treatments.
This effort is certainly not a silver bullet in the quest for solving the cancer disease, but it is a modest step in the right direction.
What is the license of this code
It is MIT.
The MIT License is a permissive free software license.
As a permissive license, it puts only very limited restriction on reuse.
The MIT license permits reuse within proprietary software provided that all copies of the licensed software
include a copy of the MIT License terms and the copyright notice.
Are there hidden strings? Will it be a commercial product?
No, we are a not-for-profit. There is no plan to commercialize this service.
Our goal is really to democratize this kind of medical benefit .
It is not important if we succeed or if others are the credited as long as that patients can have a new treatment option.