One of our New Year resolutions here at BenchSci is to share our journey with the scientific community. We decided to start with a monthly blog post that will summarize what we have recently been up to.
So here we go…
On January 10th 2017, we officially launched the Beta version of BenchSci at 10 universities across North America.
These universities include: University of Toronto, University of Calgary, City University of New York System (CUNY), State University of New York (SUNY), MD Anderson Cancer Center, Baylor College of Medicine, University of Iowa, University of Virginia, University of Wisconsin-Madison, Georgetown University, University of Notre Dame, and Thomas Jefferson University.
So far, it has been extremely successful and we are getting great feedback from our users. We conducted over 20 interviews this month with research scientists who gave us great ideas for new features (BTW, this is a great segway to our next update).
We introduced a few cool new features this month:
Explore BenchSci button – This button, which appears on the home page, allows you to view all the figures, vendor images and products that we have on the platform without entering a protein name first. This means you can filter figures by tissue, species, cell line, cell type and disease without selecting a specific protein.
New view options - We added 3 new view states which enable you to slice and dice the data however you may please! You can now view only publication figures, only vendor provided images and only products. If you prefer, you can also view all of them together.
Company filter – This new filter allows you to only view figures and additional data from your favorite vendor.
Published figures - Every day our team works on decoding more scientific papers and adding additional antibody validations in the form of figures to the platform. This month, we were able to decode additional 500,000 papers and add 100,000 new antibody validations! That being said, the current data set on BenchSci is far from complete. Our goal is to add 100,000 new antibody validations every month. This means that even if you can’t find your favorite protein on the platform now, we are probably going to add it in the next couple of weeks, so don’t forget to come back and check!
Products – In order to find papers that cite commercial antibodies, we had to partner with many antibody vendors and collect their antibody catalogs. This enabled us to create the world’s largest database of commercially available antibodies! Our current database contains 3,445,730 unique antibodies. This month we uploaded this data to the platform. This means that, if you prefer, in addition to viewing figures you can also find products that fit those rare use cases that haven’t been validated in the literature. All products are ranked based on the number of published figures that match your preferred technique.
This month we were able to improve our machine learning algorithm to achieve 90% accuracy! What does this mean? It means that now a computer can read a scientific paper like a scientist and understand the context in which an antibody was used with a 90% accuracy level.
We would say that this is “pretttttttttty pretttttttttty pretttttttttty good”.
New team members
We added two amazing engineers to our team, Anvar Gazizov and Leo Rotta-Rossi to help us develop new awesome features super-fast!
Please reach out if you have any questions, comments or ideas for new features. You can reach me at firstname.lastname@example.org