I have no idea how close we are to done but I can commment on the real time thing. I recall a member mentioning "real time" in a post and I remember thinking it was not the best term to use. Perhaps we are thinking of the same post? If so then I believe the poster was trying to express the notion that the HDC project could yield results that have a more immediate benefit as opposed to the other projects which are laying a foundation for future research with benefits to come in the more distant future.

Just to refine that answer a little: the tissue samples are from real patients, but they are from patients where the outcome (and long-term history) is known, so the scientists can take the cancer type, the treatment and the outcome, and draw conclusions about which treatment worked best for which precise sub-type of cancer. Hopefully, if this goes well, it will rapidly turn into a diagnostic tool used to determine the best treatment course for cancer patients now.

" they are from patients where the outcome (and long-term history) is known,"

That would indicate to me that there are a finite number of samples (how many). From other posts I know that they can scan 4,000 per day for us to process but it doesn't tell us the total; just that we can keep up with the 4,000 added each day.

Is our processing trying to determine:
1) the cancer type/subtype
2) an automated way of determining the cancer type/subtype
3) new cancer types and/or subtypes that are too subtle for human observation

My reading of the description is that the scientists behind this project have already found a way of automating the processing of cancer samples to evaluate known biomarkers. The role of the World Community Grid computers is to crunch the information coming in to determine the presence and levels of different biomarkers and to help find how these biomarkers cluster to improve scientist's knowledge of different cancer subtypes. The scientists will be able to use this knowledge produced by the World Community Grid to analyze how different treatments affect outcomes.

This knowledge won't spead up diagnosis so much as match treatments to patients and this can matter a great deal. All cancer drugs have pretty serious side effects, but they are used when the benefits outweigh the risks. However, not all patients with a cancer benefit from the same drugs. A decade ago, many doctors gave all their breast cancer patients tamoxifen, but then it was found that around 30% of patients had a trait that meant that they had no benefit at all. As a result of that research, doctors stopped giving tamoxifen to that group and they no longer needlessly suffer from the side effects of the drug, which include a slightly increased risk of endometrial cancer. More recently, scientists have identified a relatively small group of patients with a trait than means that tamoxifen may actually fuel the cancer in the long term, so many doctors now use different drugs for that group.

Knowing biomarkers can also help with drug development. There is a really bad type of breast cancer that only affects about 2% of breast cancer patients in the US, but is rising in incidence for unknown reasons and there has been no success in the past decade in lowering the death rate. A drug company developed a drug that targeted certain biomarkers and these just happen to be common in this small, very lethal subtype. So it conducted international research on patients with this subtype, including in Tunisia where the prevalence is high, and the drug seems to work really well. Unfortunately, only half of patients with this really bad subtype have the ideal biomarkers for benefitting from this drug, so more work needs to be done.

Please correct me if I have said anything wrong, since I must admit that I found the project description hard reading, despite knowing a fair bit about cancer and being a native English speaker.

The investigators are using a set of new and existing arrays to perform the analysis. So far, the team has digitized and forwarded more than 65, 000 work units to WCG for processing and have been generating
an average of about 4, 000 additional work units per day. There are also
new arrays which are being prepared for these experiments.

The results of the most computationally intensive operations are forwarded to the research team for further processing at which time the
output generated by the WCG will be compared with the clinical profiles of record for these patients. The WCG stages of processing are "real-time", but the later stages of analysis are much slower.