SciencePediaIn the intricate world of X-ray crystallography, determining the precise three-dimensional arrangement of atoms within a molecule is the ultimate goal. However, a fundamental obstacle known as the "phase problem" stands in the way: while we can measure the intensity of diffracted X-rays, the crucial phase information is lost, preventing a direct calculation of the molecular structure. This article explores a brilliant solution to this conundrum: the Patterson map. Developed by Arthur Lindo Patterson in 1934, this mathematical tool transforms experimental data into a "ghost map" not of atoms, but of the vectors connecting them.
We will first delve into the Principles and Mechanisms of the Patterson function, uncovering how this map of interatomic vectors is constructed and interpreted, with a focus on its key features and the power of Harker sections. Subsequently, in Applications and Interdisciplinary Connections, we will explore how this foundational concept is practically applied, from the classic heavy-atom method and modern anomalous scattering techniques to the powerful method of molecular replacement and its role as a diagnostic tool in structural biology. By the end, you will understand how the Patterson map provides the essential first foothold for solving complex molecular puzzles.
Imagine you are an archaeologist who has discovered a remarkable new artifact: a sealed, opaque box. You’re not allowed to open it, but you are allowed to use a special kind of sonar. You can send a pulse in from any direction and record the echoes that come back. You find that the echoes are very strong, which tells you there’s something substantial inside, but the detector you have is a bit primitive: it only records the intensity of the echo, not the precise timing or shape of the returning wave. This is the heart of the "phase problem" in X-ray crystallography. Our detectors measure the intensity of thousands of diffracted X-ray spots, which gives us the amplitude of the underlying waves, but all a-priori information on their relative phases is lost.
So, if we can't directly reconstruct a map of the object—the electron density —what can we do? A physicist named Arthur Lindo Patterson had a brilliant insight in 1934. He asked, "What happens if we just go ahead and do the standard mathematical reconstruction (a Fourier transform) using the data we do have—the intensities themselves?" The result, he discovered, is not a map of atoms. It’s something else, something strange but wonderfully useful. It's a ghost map, a map of echoes. This is the Patterson map, or Patterson function.
The Patterson function, , is mathematically the autocorrelation of the electron density. That sounds rather formal, so let’s unpack it. An autocorrelation is a way of measuring how similar a function is to a shifted version of itself. Imagine you have a drawing of a constellation. You make a transparent copy of it, lay it on top of the original, and then slide the copy around. Every time a star on the copy lines up with a star on the original, you make a mark at the spot corresponding to your slide vector. The resulting collection of marks is the autocorrelation.
The Patterson map is exactly this, but for the electron density of the crystal. It doesn't show you where the atoms are. Instead, it shows you the vectors between the atoms. A peak at a position in Patterson space means that if you start at some atom in the crystal and travel along the vector , you will land on another atom. The height of the peak tells you how many pairs of atoms in the unit cell are separated by that specific vector.
To start interpreting this map of vectors, we need to understand its key features.
The first thing you’ll notice on any Patterson map is an enormous peak right at the origin, where . This peak is always the largest one, but for solving the structure, it's completely useless. What is it? It's the result of every atom being compared with itself. The vector from an atom to itself has zero length. So, every single atom in the unit cell contributes to this one giant peak. It’s the sum of all the "self-vectors," and its magnitude is proportional to the sum of the squares of the scattering power of all atoms, something like , where is the atomic number of the -th atom. It's a fundamental feature, but we must look past it to find the interesting information.
The other peaks, the ones away from the origin, are what we're after. These correspond to vectors between different atoms. And here’s the clever part: the height of a Patterson peak at vector is proportional to the product of the atomic numbers of the two atoms involved, .
Let’s think about this with a simple molecule like carbonyl sulfide, OCS. The atomic numbers are , , and . The Patterson map will have a peak for the C-to-S vector whose height is proportional to . It will also have a peak for the O-to-S vector with a height proportional to . The ratio of their heights would be . This simple rule is our foothold. If we can introduce a "heavy atom" like mercury () or uranium () into a protein crystal (mostly made of C, N, O with ), the vectors between those heavy atoms will produce peaks that are vastly stronger than any others. The Patterson map will light up with these "heavy-heavy" vectors, while the thousands of "light-light" atom vectors fade into a noisy background.
One last, crucial property. Every Patterson map is centrosymmetric, meaning that for every peak at vector , there is an identical peak at . This is because if there's a vector from atom A to atom B, there is necessarily a vector from B to A, which is just the opposite vector. This holds true even if the crystal structure itself does not have a center of symmetry. The act of measuring only intensities and losing the phases imposes this symmetry upon our vector map.
So, we have a map of vectors, dominated by the bright signals from a few heavy atoms we’ve added. How do we work backward from this vector map to find the actual coordinates of those heavy atoms? This can be a tricky puzzle, especially if you have more than one heavy atom creating multiple overlapping sets of vectors.
Here, the inherent symmetry of the crystal comes to our rescue. Crystal structures are not just random assortments of atoms; they are built by repeating a single motif according to a strict set of symmetry rules, defined by the crystal's space group. Let’s say a space group has a symmetry operation that takes an atom at and creates an identical one at a new position, say . This is the rule for the common space group .
Now, consider the vector between this pair of symmetry-related atoms. Its coordinates will be:
Look closely at that vector! No matter what the original coordinates of the atom were, the component of the vector between it and its symmetry mate is always , which in the periodic world of a crystal is the same as . This is a fantastic revelation! It means that all the strong peaks corresponding to vectors between symmetry-related heavy atoms are not scattered randomly throughout the 3D Patterson map. They must all lie on a specific 2D plane: the plane where .
These special planes (or lines, for other symmetries) where symmetry-related vectors are forced to lie are called Harker sections. They are the key to solving the puzzle. Instead of a bewildering 3D search for our heavy-atom peaks, we can limit our attention to a simple 2D slice of the map, dramatically reducing the complexity of the problem. By locating a few consistent peaks on these sections, we can deduce the coordinates of the heavy atom that created them.
The Patterson method is not a panacea. If a crystal structure has two different heavy atoms, say A and B, the map becomes a superposition of three patterns: vectors from A to its symmetry mates, vectors from B to its mates, and a third set of "cross-peaks" from A to B. Unscrambling this becomes a much harder puzzle.
Worse, what if your molecule of interest doesn't have a convenient place to put a very heavy atom, or all its atoms have similar atomic numbers? The Patterson map becomes a crowded, confusing jumble of thousands of similarly-sized peaks.
Here, modern crystallography employs an even more subtle trick using anomalous scattering. By carefully tuning the energy of the X-rays to be near an "absorption edge" of a particular element (say, Selenium, which can often be substituted for Sulfur in proteins), we can make that element's scattering behavior change. We can collect data at two energies, one "on" the edge and one "off" it. If we then calculate a Patterson map for each dataset and subtract one from the other, an amazing thing happens. Most of the background mess cancels out. What remains is a clean map showing only vectors that involve the atom whose edge we tuned to. This "difference Patterson" technique allows us to pick out specific atomic vectors even in the most crowded of environments.
From a seemingly useless calculation on incomplete data, the Patterson map emerges as a beautiful and powerful tool. It transforms the intractable phase problem into a solvable geometric puzzle, allowing us to find that crucial first foothold on the long climb toward seeing the atomic structure of a molecule.
We have now acquainted ourselves with the curious beast that is the Patterson function. We understand its nature: a map not of atoms, but of the vectors that connect them, born from the raw intensities of a diffraction experiment, seemingly barren of the phase information we so desperately need. It might appear to be an arcane mathematical detour, a map of nowhere particularly useful. But to the working scientist, this vector map is a treasure chart. It is a Rosetta Stone for the language of crystals, a diagnostic tool for structural maladies, and a powerful engine for assembling the atomic world. Let us now embark on a journey to see how this abstract concept springs to life, solving real problems across the landscape of science.
The most celebrated role of the Patterson map is that of a detective's aid in the hunt for atomic structure. The core challenge in crystallography, the phase problem, is like being in a room filled with echoes but with no sense of where the sounds originated. The Patterson function was the first systematic method developed to begin pinpointing those sources. The strategy, in its classic form, is one of brilliance and cunning: introduce a "heavy atom" into the crystal.
Imagine a protein crystal, a metropolis of thousands of light atoms like carbon, nitrogen, and oxygen. Now, we introduce a single, ponderous atom—a mercury or a gold—into each molecule. This heavy atom, with its large cloud of electrons, a veritable giant in a city of commoners, scatters X-rays far more strongly than its neighbors. It acts as a beacon. While we still can't see the atoms directly, the influence of this beacon is stamped all over the diffraction pattern.
Here is where the first piece of Patterson's magic comes into play. By collecting diffraction data from both the native crystal and the heavy-atom derivative, we can compute a difference Patterson map. This map is generated not from the total intensities, but from the squared difference between the derivative and native amplitudes. The logic is one of elegant subtraction: the city of light atoms is common to both crystals, so its contribution largely cancels out. What remains are the new echoes, the new vectors, created by the beacon itself. The resulting map is wonderfully simplified, primarily showing only the vectors between the heavy atoms in the unit cell. The haystack has been cleared away, leaving only the needles.
But the map shows vectors, not positions. If our crystal unit cell contained just two atoms, one at the origin and one at an unknown position , the Patterson map would happily show us a peak at , directly revealing the second atom's location. But in a real crystal, the heavy atom could be anywhere. So how do we convert a vector into a position?
The answer lies in a beautiful piece of synergy: we use the crystal's own inherent symmetry against itself. Any crystal, by definition, possesses symmetry operations that relate the positions of atoms. A simple rotation, for instance, might dictate that if there is an atom at , there must also be one at . The vector connecting this pair of symmetry-related atoms is therefore . This is no ordinary vector. Its form is strictly determined by the symmetry operation. All such vectors, for any atom in the crystal, will fall onto a specific line or plane in the Patterson map—a special region known as a Harker section.
This is a profound simplification. Instead of searching the entire three-dimensional volume of the Patterson map for clues, we need only look at these specific, symmetry-defined Harker sections. Finding a peak at on the Harker plane immediately tells us the and coordinates of our heavy atom! The crystal's symmetry has been leveraged to decode the vector map. In more complex cases, multiple symmetry operations provide multiple Harker vectors for the same atom, allowing for a cross-check that builds a unique and robust solution. The puzzle can become quite intricate and satisfying. For a centrosymmetric crystal, where an atom at implies a partner at , the Patterson map will contain not only the vector from a central heavy atom to the atom , but also the Harker vector between the atom and its symmetric mate. A crystallographer can scan the list of peaks for pairs that fit this ratio, and each such pair unmasks the location of another atom in the structure.
The classic heavy-atom method involved physically soaking crystals in heavy-metal solutions. Modern structural biology often uses more subtle approaches, and the Patterson concept has evolved in lockstep. Many proteins are now produced with methionine residues replaced by selenomethionine. The selenium atom is a modest "heavy atom," but its scattering behavior has a special property: it changes dramatically as the X-ray wavelength is tuned across its absorption edge. This phenomenon is called anomalous scattering.
This opens the door to crafting even more specific Patterson maps. By collecting data at different wavelengths, or by comparing the intensities of reflections that should be equivalent by symmetry (so-called Friedel pairs), we can again isolate the signal from just the selenium atoms. One can compute an anomalous difference Patterson from the differences in Friedel pairs at a single wavelength, or a dispersive difference Patterson from the differences between data at two different wavelengths. Both maps reveal the network of vectors between the selenium atoms, but they do so by tapping into different physical components of the anomalous scattering signal. They are like two different stains that both highlight the same target structure, providing complementary and confirmatory evidence for the selenium substructure, which is the first step toward phasing the entire protein.
So far, we have been hunting for a few well-placed beacons. But what if we already have a good idea of what the molecule looks like? Perhaps we have solved the structure of a similar protein from another species. This is the starting point for a profoundly different and powerful application of the Patterson function: Molecular Replacement.
The problem now is not to find a few atoms, but to find the orientation and position of an entire known molecule (the "search model") within the new crystal's unit cell. How can the Patterson map, a seemingly jumbled collection of all interatomic vectors, possibly help? The key insight is that the Patterson map contains two types of vectors: intramolecular vectors (between atoms within the same molecule) and intermolecular vectors (between atoms in different molecules). The intramolecular vectors depend only on the molecule's structure and orientation, not its position in the unit cell. This set of internal vectors forms a unique fingerprint of the molecule.
The first step of molecular replacement is therefore a rotation function. It is a systematic search to find the correct orientation of the search model. Computationally, we can imagine taking our search model, calculating its own internal vector map (its theoretical Patterson function), and then "tumbling" it through all possible three-dimensional orientations. At each orientation, we ask: "How well does the fingerprint of my model's vectors match the features in the experimental Patterson map?" The orientation that yields the highest overlap, the best correlation score, is declared the winner.
This process can be visualized by considering a simpler case: finding the orientation of a known, rigid cofactor, like a flat, square-shaped metal-porphyrin ring, within a larger protein. We know the vectors from the central metal to the four corners of the square. We can then computationally place this vector "star" onto our experimental Patterson map and rotate it, summing up the Patterson density values at the tips of the vectors. The rotation angle that places the vector tips onto the strongest Patterson peaks is the most likely orientation of the cofactor in the crystal. Once this rotation is found, a second step—the translation function—finds its correct position. But it is the Patterson-on-Patterson comparison that cracks the first half of the problem.
Perhaps the most surprising and subtle applications of the Patterson map are in diagnostics. A scientist must not only use their tools to get the right answer but also to understand when something has gone wrong. The Patterson map is an exquisite tool for revealing hidden pathologies in a crystal.
A classic example is translational non-crystallographic symmetry (tNCS). This occurs when the crystal's asymmetric unit—the fundamental building block—contains more than one copy of the molecule. If two identical molecules are related by a simple shift, a translation vector , then every atom in the first molecule has a counterpart in the second, separated by the exact same vector . The result in the Patterson map is the superposition of thousands of identical vectors, all adding up to create a single, enormously strong peak at the position corresponding to . Finding such a strong, non-Harker peak is the unambiguous signature of tNCS. It's a flashing red light, alerting the crystallographer that the problem is more complex and must be handled with special care.
An even more vexing problem is pseudosymmetry, where the crystal lattice or the arrangement of molecules almost possesses a higher symmetry than the true space group. A monoclinic crystal with its unique angle happening to be very close to , for example, will diffract as if it were orthorhombic. In this case, the Patterson map can become a nightmare of ambiguity. The expected clean peaks from a heavy-atom substructure can be lost in a forest of "ghost" peaks generated by the near-symmetry. The map appears to have a higher symmetry than it should, and the true vector set is superimposed with a ghost copy of itself, making it nearly impossible to interpret. While frustrating, this is invaluable information. The Patterson map is telling us that our simple assumptions are wrong and that a more sophisticated model is needed to untangle the structure.
Finally, the very appearance of the map is a direct report on the quality of the experiment itself. A Patterson map calculated from low-resolution data is like a blurry photograph. The dominant Harker peaks may appear as large, diffuse blobs, but the vast majority of interatomic vectors are smeared together into an unresolved, noisy background. In contrast, a map from high-resolution data is a sharp, detailed image. The Harker peaks become tight and well-defined, and the background resolves into a rich, complex tapestry of thousands of individual, weaker peaks. This visual transformation is a potent reminder that our ability to see the atomic world is limited only by the quality of our observation.
From a simple set of vector rules, the Patterson function has blossomed into a field of scientific inquiry all its own. It is the crystallographer's compass, puzzle box, and magnifying glass, revealing not only the elegant architecture of molecules but also the subtle imperfections that make the world of real crystals so challenging and fascinating. It is a stunning example of how a deep, physical intuition, married to a simple mathematical idea, can grant us vision into a world once thought to be invisible.