Summary: DeepMind's AlphaFold algorithm for predicting the 3D shape of protein folding is one of the biggest scientific breakthroughs in biology the last 50 years and solves a long standing problem. The breakthrough will likely lead to the Nobel Prize in chemistry and advance drug discovery on a scale we could only dream of a decade ago. AlphaFold underpin why our NextGen Medicine basket is such an important theme for investors to be exposed to long-term as biotechnology companies could end up being the big winners over the next couple of decades.
DeepMind’s gift to humanity will unleash massive innovation in medicine
In the midst of a global pandemic the world experienced two major breakthroughs in medicine and biochemistry in November 2020. First, Pfizer with its partner BioNTech announced a highly effective vaccine against Covid-19 using mRNA technology for the first time ever in history; it became the fastest developed vaccine in the world. Later during the month Nature announced a ground-breaking scientific discovery from Google’s AI unit DeepMind in which the company’s deep learning algorithm called AlphaFold is able to predict the 3D shape of proteins. Scientists agreed that “It will change everything” and DeepMind chose to make their AlphaFold Protein Structure Database available for free to all scientists.
Around 20 months later DeepMind went a step further announcing that it has, in collaboration with the European Bioinformatics Institute, catalogued nearly all protein known in the universe expanding the database of protein structures from 1 to 200 million structures. Before AlphaFold experimental work had unveiled around 190,000 protein structures. This new database gifted to humanity will advance our understanding of biology and potentially lead to quantum leaps in drug discovery. DeepMind’s scientific breakthrough solves the 50-year problem of protein folding and will most likely lead to the Nobel Prize in Chemistry.
Investors must have exposure to “biology” in their long-term portfolios
The 21st century has often been labelled the century of biology in which computational biology and breakthroughs such as AlphaFold will deliver better and faster drugs to combat diseases. Our NextGen Medicine basket consists of 30 companies representing the long tail of innovation inside the biotechnology industry which today is dominated by a group of very large and successful companies such Amgen, Gilead Sciences, and Vertex Pharmaceuticals etc. Our focus has been on some of the newer technologies within biotechnology such as mRNA, gene therapy, gene editing etc.
The world’s leading biotechnology index which tracks the largest and most liquid biotechnology companies is called NASDAQ Biotechnology Index and it has returned 9.7% annualised since late 2003 compared to 8.1% for MSCI World Index generating 1.6% annualised alpha. The amount of alpha has shrunk considerably the past year as rising interest rates has pushed down equity valuations of high duration assets (companies with high valuation or no earnings) and made financing more difficult.
With the type of drug discovery that might derive from AlphaFold and increased understanding of proteins will make biotechnology as massive and very profitable trend over the coming decades. Microsoft and the computer industry have generated 5-10% annualised alpha since the late 1980s and we believe biotechnology will ultimately do the same catapulting some of these companies into the top ranks of equity indices in terms of market value. One can invest in the stocks in our theme basket but requires decent amount of due diligence and it comes with the risks of not choosing the winners. An alternative is ETFs such as the iShares Nasdaq Biotechnology UCITS ETF (2B70:xetr) tracking the main benchmark index within biotechnology or the Global X Genomics & Biotechnology UCITS ETF (GNOM:xmil) which closer mimics our NextGen Medicine basket.
| Name | Mkt Cap (USD mn.) | Sales growth (%) | R&D in % of mkt cap | Diff to PT (%) | 5yr return |
| Moderna Inc | 72,955 | 227.6 | 3.3 | 6.1 | NA |
| BioNTech SE | 43,313 | 831.0 | 2.4 | 33.9 | NA |
| Illumina Inc | 35,049 | 34.0 | 3.7 | 35.0 | 14.7 |
| Seagen Inc | 32,274 | -25.4 | 4.2 | 4.6 | 256.0 |
| Genmab A/S | 23,386 | -16.5 | 19.2 | 2.7 | 87.0 |
| Argenx SE | 20,279 | -74.4 | 1.9 | 9.5 | 1,930.3 |
| Exact Sciences Corp | 8,422 | 13.3 | 4.4 | 49.7 | 23.0 |
| Grifols SA | 8,336 | 0.1 | 4.3 | 58.4 | -38.5 |
| Swedish Orphan Biovitrum AB | 6,751 | 21.0 | 32.9 | 12.6 | 84.7 |
| CRISPR Therapeutics AG | 6,087 | 83,039.3 | 7.7 | 45.1 | 327.1 |
| Ionis Pharmaceuticals Inc | 5,817 | 18.9 | 11.4 | 18.7 | -21.7 |
| Intellia Therapeutics Inc | 5,439 | -26.5 | 6.0 | 78.5 | 345.3 |
| 10X Genomics Inc | 4,807 | 50.0 | 4.9 | 79.8 | NA |
| Natera Inc | 4,718 | 48.5 | 6.5 | 63.1 | 491.0 |
| Arrowhead Pharmaceuticals Inc | 4,690 | 195.6 | 5.7 | 72.8 | 2,310.9 |
| Beam Therapeutics Inc | 4,428 | 251,025.0 | 5.9 | 46.4 | NA |
| Denali Therapeutics Inc | 4,083 | -75.6 | 7.1 | 105.9 | NA |
| Mirati Therapeutics Inc | 3,992 | 495.3 | 13.3 | 61.7 | 1,395.4 |
| Ultragenyx Pharmaceutical Inc | 3,754 | -7.0 | 14.2 | 109.0 | -15.3 |
| Abcam PLC | 3,573 | 20.3 | 0.9 | 17.7 | 27.4 |
| Galapagos NV | 3,486 | 3.6 | 13.2 | 20.7 | -19.2 |
| PTC Therapeutics Inc | 3,411 | 32.3 | 16.0 | 4.3 | 142.1 |
| Fate Therapeutics Inc | 3,183 | 42.2 | 8.7 | 129.9 | 1,016.0 |
| Twist Bioscience Corp | 2,589 | 44.3 | 3.6 | -9.8 | NA |
| Sage Therapeutics Inc | 2,140 | -99.4 | 14.7 | 64.4 | -58.6 |
| Allogene Therapeutics Inc | 1,956 | -99.5 | 11.5 | 98.2 | NA |
| Iovance Biotherapeutics Inc | 1,892 | NA | 14.9 | 126.2 | 115.0 |
| CareDx Inc | 1,304 | 39.4 | 6.3 | 102.7 | 1,601.4 |
| Pacific Biosciences of California Inc | 1,169 | 31.9 | 14.8 | 127.1 | 29.9 |
| Invitae Corp | 456 | 50.6 | 101.7 | 181.4 | -80.8 |
| Aggregate / median | 323,738 | 31.9 | 6.8 | 54.0 | 87.0 |
Why is the protein folding problem so important?
In a recent podcast interview with Lex Friedman, Demis Hassabis (co-founder of DeepMind) describes why protein folding is such an important problem to solve. Proteins are the building blocks of life and our bodies and are responsible for most of what happens inside cells. Drugs for curing a disease attach to certain proteins and therefore you have to know their 3D shape to design the drug properly. With AlphaFold biologists can now look up the same of a protein in the database instead of using eight years on average, in some cases scientists have given up on finding certain protein structures, to determine the structure. It will turbocharge drug discovery and protein design.
The graphics below show how much better AlphaFold is relative to the competition in predicting protein structures with a score above 90 being on par with experimentally determined structures.
Source: NextGen Medicine basket and the protein folding problem | Saxo Group (home.saxo)
